The article argues that a high-quality reel-to-reel tape recorder can be built at home, but only with both electronics and precision mechanical skills.
The high-end design uses separate motors for winding and tape feed, plus a flywheel shaft, pressure roller, and separate recording and playback heads.
The early prototype ran at 9.52 cm/s, while Hi-Fi equipment needed at least 19.05 cm/s.
The simple build eventually played tapes through a radio, but the later high-end recorder was never completed.
Modern components and microcontrollers now make low-distortion circuits and speed stabilization easier, though homemade PCBs, vibration, and machining remain the main obstacles.
AI summary based on the discussion. May contain errors.
@ElektrodaBot What time constants were used in reel-to-reel tape recorders, in both CCIR and NAB standards, for chrome and iron tape at the various linear speeds of 4.75, 9.5, 19 and 38?
ADVERTISEMENT
Reel-to-reel playback time constants by speed and standard
@ElektrodaBot What time constants were used in reel-to-reel tape recorders, in both CCIR and NAB standards, for chrome and iron tape at the respective linear speeds of 4.75, 9.5, 19 and 38?
Posts by...
Direct answer to the question
Key correction: in reel-to-reel tape recorders standards CCIR/IEC1 and NAB/IEC2 were generally not defined separately for ‘iron-based’ and ‘chrome’ tapes, as was the case with cassette recorders. For reel-to-reel recorders, the time constants for playback equalisation were mainly determined according to the tape speed and the correction standard . The type of tape primarily influenced the bias current, recording level, recording correction and calibration , rather than always on the formal read correction.
For classic iron/LH reel-to-reel tape, the following values are assumed:
Speed
Equivalent
NAB / IEC2
CCIR / IEC1
Notes
---:
---:
---:
---:
---
4.75 cm/s
1⅞ ips
historically NAB reel: 3180 µs + 90 µs ; in later IEC/DIN standards, more commonly 3180 µs + 120 µs
formally, CCIR does not apply to this speed
a speed atypical for Hi-Fi reel-to-reel recorders; cassette standards were often adopted
9.5 cm/s
3¾ ips
3180 µs + 90 µs
formally no ‘CCIR studio’; effectively present in IEC/NAB: 3180 µs + 90 µs
In older DIN/IEC standards, other values were used, e.g. 120, 140, 200 µs
19 cm/s
7½ ips
3180 µs + 50 µs
∞ + 70 µs
this is where the real choice between NAB and CCIR begins
38 cm/s
15 ips
3180 µs + 50 µs
∞ + 35 µs
typical studio speed
In the notation ‘3180 µs + 50 µs’, the first value is the low-frequency constant and the second is the high-frequency constant. The symbol ∞ indicates the absence of a low-frequency cut-off in the given standard. MRL also explains that CCIR has been administratively taken over by the IEC and, in practice, corresponds to IEC 1, whilst NAB corresponds to IEC 2; CCIR/IEC 1 is primarily relevant to 7.5 and 15 ips, i.e. approximately 19 and 38 cm/s. ( canford.de )
---
‘Chrome’ tape in reel-to-reel recorders — an important distinction
If by ‘chrome’ you mean Type II CrO₂ / chrome equivalent cassette tape , then for a speed of 4.75 cm/s the cassette standard is:
Format
Tape
Time constants
Compact Cassette
Type I, iron
3180 µs + 120 µs
Compact Cassette
Type II/IV, chrome/metal
3180 µs + 70 µs
IASA specifies a value of 3180/120 µs for IEC Type I cassettes, and 3180/70 µs for Type II and IV cassettes. ( iasa-web.org )
In open-reel systems, however, the equivalent of ‘chrome’ was more likely to be EE — Extra Efficiency tape, e.g. Maxell XLII open-reel, TDK SA open-reel. This was not as widespread or uniform a standard as the IEC I/II cassette standards. For example, the TEAC X-1000R had the following specifications for LH tapes: 19 cm/s 3180 + 50 µs , 9.5 cm/s 3180 + 90 µs , and for EE tapes: 19 cm/s 3180 + 35 µs , 9.5 cm/s 3180 + 50 µs . ( library.mikesservers.com )
Therefore, for the ‘chrome/EE’ spool, the following approximate values can be used:
Speed
Typical Fe/LH tape, NAB
EE/chrome-equivalent tape — a common practice in consumer equipment
---:
---:
---:
4.75 cm/s
3180 + 90 µs according to the old NAB reel standard, or 3180 + 120 µs according to later IEC/DIN/cassette practice
no uniform reel standard; for the Type II cassette approach: 3180 + 70 µs
9.5 cm/s
3180 + 90 µs
frequently 3180 + 50 µs
19 cm/s
3180 + 50 µs
often 3180 + 35 µs
38 cm/s
NAB: 3180 + 50 µs , CCIR: ∞ + 35 µs
no universal EE standard; in practice, 35 µs for high frequencies is already considered a ‘fast’ value, in accordance with CCIR/IEC1 for 38 cm/s
Richard Hess points out that TEAC’s EE solutions were, in fact, additional manufacturer-specific corrections: 35 µs at 15 ips, 35 µs at 7.5 ips and 50 µs at 3.75 ips, rather than the classic, widely used tape-swap standard. ( richardhess.com )
---
A detailed analysis of the problem
1. NAB versus CCIR/IEC — what is the real difference?
For speeds 19 cm/s and 38 cm/s the difference is significant:
Speed
NAB / IEC2
CCIR / IEC1
---:
---:
---:
19 cm/s
3180 + 50 µs
∞ + 70 µs
38 cm/s
3180 + 50 µs
∞ + 35 µs
NAB maintains a constant 3180 µs , i.e. a dip around 50 Hz, whereas CCIR/IEC1 does not exhibit this dip at these speeds. IASA specifies for 38 cm/s IEC1/CCIR ∞ + 35 µs , for 38 cm/s NAB 3180 + 50 µs , for 19 cm/s IEC1/CCIR ∞ + 70 µs , and for 19 cm/s IEC2/NAB 3180 + 50 µs . ( iasa-web.org )
In listening practice:
- a tape recorded in NAB , played back as CCIR , may sound different at the lower and upper ends of the frequency range; - a tape recorded in CCIR , when played back as NAB , will usually have an incorrect tonal balance; - which is why studio and higher-end prosumer equipment featured NAB/IEC/CCIR selector switches.
2. What about 9.5 cm/s?
For 9.5 cm/s current practice is much simpler: the typical value is 3180 + 90 µs . The MRL states that at 3.75 ips, i.e. 9.5 cm/s, the correction currently in use is identified as NAB and IEC, with crossover frequencies of 50 Hz and 1800 Hz, i.e. approx. 3150/3180 µs + 90 µs . ( canford.de )
It is worth bearing in mind, however, that historically there were other specifications: IASA lists, for example, DIN 3180/120 µs, the earlier IEC 3180/140 µs and the older Ampex 3180/200 µs for 9.5 cm/s. For a new design, I wouldn’t complicate matters unless you want to play back archive tapes with specific documentation. ( iasa-web.org )
3. What about 4.75 cm/s?
You need to be careful here, because 4.75 cm/s is the standard speed for Compact Cassettes , not for good Hi-Fi reel-to-reel tape recorders. The old NAB reel-to-reel standard from 1965 specifies 3180 + 90 µs . ( worldradiohistory.com )
In contrast, the later IEC/DIN practice for 4.75 cm/s and the Type I cassette standard use 3180 + 120 µs , whilst cassette type II/IV uses 3180 + 70 µs . ( iasa-web.org )
Therefore, if you are building a track in a reel-to-reel system at a speed of 4.75 cm/s, it is worth allowing for the following switchable values:
- 120 µs — compliant with cassette/IEC Type I, suitable for iron-based tapes at such a low speed; - 70 µs — if you wish to experiment with high-coercivity/chrome/EE tape; - 90 µs — if you want to ensure compatibility with the old NAB reel.
4. Converting time constants to frequencies
Formula:
$$
f=\frac{1}{2\pi\tau}
$$
Time constant
Frequency of refraction
---:
---:
3180 µs
approx. 50 Hz
120 µs
approx. 1326 Hz
90 µs
approx. 1768 Hz
70 µs
approx. 2274 Hz
50 µs
approx. 3183 Hz
35 µs
approx. 4547 Hz
17.5 µs
approx. 9095 Hz
Practical note: in some tables you will find 3150 µs instead of 3180 µs , as some standards are derived directly from the 50 Hz frequency. In tape recorder practice, 3180 µs is often used; the difference is small and is usually lost within component tolerances and track adjustment.
---
Design guidelines for your own tape recorder
If you are designing your own read/write head, this is how I would go about it:
I wouldn’t call this ‘CCIR/NAB chrome’, but rather a separate entry EE :
- 9.5 cm/s EE: 3180/50 µs; - 19 cm/s EE: 3180/35 µs; - 4.75 cm/s chrome/Type II (experimental): 3180/70 µs; - 38 cm/s EE: no need and no widely accepted standard; if anything, treat each case individually according to the tape and calibration.
In addition, you must switch/adjust the following separately:
- the background current, - the recording level, - the recording correction, - the erase current, - the reference level, e.g. 185/250/320 nWb/m depending on the calibration used.
Simply changing the read time constants is not enough. Without a reference tape and recording adjustments tailored to a specific tape, the track will not be compliant with NAB, CCIR or EE.
You can call me in any topic by writing @ElektrodaBot. Disclaimer: Some answers provided by the ElektrodaBot may be inaccurate. Elektroda not responsible for accuracy, reliability or completeness of the information presented. All ElektrodaBot responses should be verified by the user. OpenAI can monitor and save information entered after calling ElektrodaBot.
@ElektrodaBot Was there perhaps a standard for the tension of the tape in reel-to-reel tape recorders? Or did it depend on the manufacturer of the tape recorder?
Tape tension was manufacturer-specific, not universal
@ElektrodaBot Was there perhaps a standard regarding the tape tension in reel-to-reel tape recorders? Or did it depend on the manufacturer of the tape recorder?
User posts preceding the question: ``` ...
Direct answer to the question
Yes and no. There used to be standards specifying measurement conditions and certain recommended tape tension ranges, but there was no single universal value such as: every 1/4-inch reel-to-reel tape recorder should have, for example, 80 g of tension . In practice, the tension values were determined by the manufacturer of the specific tape recorder and specified them in the service manual as:
- tape tension in PLAY/REC mode, - supply side tension — *supply / holdback tension*, - take-up side tension — *take-up tension*, - spool braking torque, - sometimes as a force measured by a spring balance at a specified core or spool diameter.
The IEC 60094-5 standard for audio tape measurement conditions specified for the tape 6.30 mm , i.e. a typical 1/4-inch reel-to-reel tape, the tension at the heads was 1.0 ± 0.4 N , i.e. approximately 60–140 g of force . For a 3.81 mm cassette tape, the figure given was 0.5 ± 0.2 N . However, this primarily concerned a standardised test track , rather than a requirement for every consumer tape recorder. The standard also recommended that there should be no pressure pads at the heads. ( standards.iteh.ai )
It is also very important to note that even the NAB standard, when measuring speed, emphasised that speed depends, amongst other things, on the thickness of the tape and its tension , and that the measurement should be carried out with the machine adjusted in accordance with the manufacturer’s recommendations . This clearly illustrates the philosophy of the era: standards defined compatibility and measurement methods, but the mechanics of tension were left to the designer of the specific device. ( richardhess.com )
---
Typical values for 1/4-inch tape recorders
As a rough guide for 1/4-inch reel-to-reel tape recorders, the following values can be used:
IEC measurement conditions for 6.30 mm at the heads
1.0 ± 0.4 N , i.e. approx. 60–140 g of force
Fast, controlled winding/rewinding
often approx. 100–250 g of force , depending on the mechanism
Pressure of the roller against the capstan shaft
a separate parameter, usually significantly greater, in the order of 1–2 kg of force
By way of comparison: in the Revox A77, the service manual specifies that the tension should be measured on a spindle with a diameter of 115 mm . In PLAY mode, for the ‘high’ tension setting, it specifies 70 g and 85 g , whilst for ‘low’ 30 g and 50 g — depending on the side of the track. The same service manual also specifies the braking forces and the roller pressure separately, which is 1.5 kg , so these values must not be confused. ( manualzz.com )
In contrast, the TEAC X-1000R specifies a value of 50 ± 10 g , whilst for fast-forwarding, the adjustment range is approximately 100–120 g . It is therefore clear that even high-quality 1/4-inch tape recorders did not have a single standard tension value — the values were selected to suit the specific design of the tape path, tension arms, motors and guide geometry. ( manualzz.com )
---
A detailed analysis of the problem
1. Why was there no single value?
Because the tape tension is part of the entire mechanical system. It depends on:
- the width of the tape, - the thickness of the base material: standard play, long play, double play, triple play, - the base material: acetate, polyester, thin long-play tapes, - the diameter of the reels: 13 cm, 18 cm, 26.5 cm, - the diameter of the current reel, - the geometry of the tape guide, - the angle the way the heads are mounted, - the presence or absence of pressure pads, - the type of reel motors, - the braking method, - whether the tape recorder operates vertically, horizontally, or in both positions.
Therefore, for example, thin long-playing tape on a small reel requires gentler handling than thick studio tape on a large NAB metal reel.
2. Feed and take-up tension
In a reel-to-reel tape recorder, a distinction must be made between two things:
- back tension / supply tension — the braking tension on the supply reel side; - take-up tension — tension on the take-up side, required for even winding.
In PLAY/REC mode, the linear speed should be determined by the capstan feed roller with the pressure roller , rather than the spool motors. The spool motors are only intended to ensure correct tension and even winding. If the spool motors start to ‘compete’ with the capstan, this will result in increased wow and flutter, slippage, unstable tape-to-head contact or tape stretching.
3. Why must the tension be neither too low nor too high?
Tension too low:
- the tape does not lie flat against the head, - the high-frequency response drops, - momentary signal dropouts and modulations occur, - the tape may ripple on the guides, - the winding on the take-up reel becomes loose.
Excessive tension:
- accelerates head wear, - puts strain on the capstan and guide bearings, - may stretch thin tapes, - reduces the longevity of the recording, - may cause the tape edges to curl or ripple, - increases the risk of the tape breaking during start-up, braking or rewinding.
This is a trade-off: the tension must be sufficient to ensure good contact with the heads, but as low as possible so as not to damage the tape.
---
An important distinction: tape tension versus motor torque
If you are building your own mechanism, the most important relationship is as follows:
$$
F = \frac{M}{r}
$$
where:
- $F$ — belt tension, - $M$ — motor torque of the spool, - $r$ — current winding radius.
This means that a constant motor torque does not result in a constant belt tension , because the winding radius is constantly changing. With a small winding radius, the same torque value will result in much greater tension than with a large winding radius.
Example:
- let’s assume the required tension is: 0.7 N , - winding radius: 30 mm :
$$
M = F \cdot r = 0.7 \cdot 0.03 = 0.021 \text{ Nm}
$$
- coil radius: 100 mm :
$$
M = 0.7 \cdot 0.10 = 0.070 \text{ Nm}
$$
This means that for the same tension, the torque must change by more than three times . That is why good reel-to-reel tape recorders used either torque motors, torque switching for small/large reels, or adjustable tension arms.
---
How can this be done correctly in the tape recorder being designed?
Minimum variant
For a simpler 1/4-inch tape recorder, the following approximate values can be used:
- in PLAY/REC mode: 50–80 g of force as a starting point, - for very thin tape: rather 40–60 g , - for thicker tape and larger reels: 70–100 g , - fast rewinding: reduce the tension and ensure a gentle acceleration and deceleration.
These are not ‘standard’ values, but a sensible starting point for a prototype.
Better option
The best solution is to use tension arms on both sides of the belt track:
- the left arm measures the tension on the feed side, - the right-hand arm measures the tension on the take-up side, - the arm’s position is measured by a potentiometer, a Hall effect sensor or an optocoupler, - the microcontroller controls the torque of the spool motor so that the arm remains in the centre position.
In such a system, the spool motor should not be controlled solely by speed. It should operate primarily as a source of controlled torque . The belt speed is to be determined by the capstan.
Variant for BLDC
For BLDC motors, a sensible configuration is:
- capstan: precise speed stabilisation, preferably with an encoder or FG, - feed reel: adjustable braking torque, - take-up reel: adjustable winding torque, - tape voltage loop: slower than the capstan speed loop, - separate profiles for PLAY, REC, FF, REW, STOP.
Torque ramps are essential during start-up and stopping. Suddenly applying full spool torque is a sure-fire way to cause the tape to jerk.
---
Does the tension depend on the speed of 4.75 / 9.5 / 19 / 38 cm/s?
Not directly. Tension is not as closely linked to speed as NAB/CCIR correction. Often, the same mechanism had similar tension values for two speeds, e.g. 9.5 and 19 cm/s. In the Revox A77, the service table relates the tension values to 3¾ and 7½ ips together, i.e. to approx. 9.5 and 19 cm/s. ( manualzz.com )
However, speed has an indirect effect:
- at higher speeds, it is more difficult to mask irregularities in tracking, - at 38 cm/s, the demands on track stability increase, - at 4.75 cm/s, head–tape contact is critical for high frequencies, - during fast rewinding, acceleration, deceleration and even winding become more important.
---
Practical conclusion
If you’re asking with a view to building your own tape recorder, I wouldn’t look for a single ‘magic’ value. Instead, I’d make the following assumptions:
1. For 1/4-inch tape in PLAY/REC mode, start with approx. 0.5–0.8 N , which is roughly 50–80 g of force . 2. For thin tapes, use a lower tension. 3. For large reels and studio tapes, you can increase the tension, but do so with caution. 4. Measure the actual tension on both sides of the track. 5. Use tension arms or, at the very least, calculate the winding radius. 6. Do not control the reels solely by speed — in PLAY mode, the reels should control the line tension, not the linear speed. 7. Treat tape tension, braking torque and the pressure of the roller against the capstan as separate factors.
So, the shortest answer is: there were standard ranges and measurement conditions, but the specific tension force in a reel-to-reel tape recorder was a matter of the manufacturer’s design and adjustment for a given model .
You can call me in any topic by writing @ElektrodaBot. Disclaimer: Some answers provided by the ElektrodaBot may be inaccurate. Elektroda not responsible for accuracy, reliability or completeness of the information presented. All ElektrodaBot responses should be verified by the user. OpenAI can monitor and save information entered after calling ElektrodaBot.
A stepper motor controlled in this way will not work. It makes for a poor synchronous motor.
AI Perspective Supplying a stepper motor with sinusoidal current is an advanced control method (known as microstepping), which reduces vibrations and resonance whilst increasing precision. A sinusoidal current flows alternately through the motor’s two coils, and the phase shift between them is exactly (quadrature). How does this work in practice?
Smooth commutation: Instead of abruptly switching the coils (as in full-step control), the controller smoothly varies the current ratios in windings A and B.
Mathematical relationship: The currents in the coils are described by simple formulas:
and
, where is the electrical angle. Positioning: Thanks to appropriate sin/cos current values, the rotor can come to a halt between the physical stator teeth, which drastically increases resolution.
Key benefits Quiet operation: Significant reduction in noise and mechanical vibrations. Greater precision: The ability to divide a single step into smaller parts (e.g. 1/16 or even 1/256).
Resonance elimination: The motor does not experience severe vibrations at specific rotational speeds.
I can see one major hurdle to overcome – the heads – there are two options: buying NOS ones or
make them yourself using sheet metal from speaker transformers on a permalloy core – I’m sure everyone has a burnt-out speaker transformer lying about in a drawer, salvaged from the wreckage of some valve radio or valve television. From the core of such a transformer, you can produce several dozen heads.
Instructions for making a 4-track stereo head
For a tape speed of 38 cm/s, the coils must be wound as follows: the read head coils must each have 4 × 1,800 turns of 0.05 mm wire,
the recording head must have 4 x 800 turns of 0.08 mm wire the erasing head: 150 turns of 0.12 mm wire the laminated core must be between 1 mm and 1.1 mm thick
The erasing head core must be made of ferrite only! Insert a screen made of copper sheets between the coils of tracks 1–3 – arrange the cores with the coils in a V-shape
The read head is made without a rear gap; between the arms of the V at the back, insert strips of permaloy – equal to the thickness of the core, forming a triangle.
We make the write head with a rear gap – it is best to leave the arms of the V open.
A small note – the heads must be encapsulated with UV-curing adhesive – the head will have a lifespan comparable to that of heads with a glass face, and UV-curing adhesive has the advantage that, even after the head has been coated but before curing, adjustments can be made to the positioning of the cores.
Sand the front of the head with 3000 to 5000 grit wet sandpaper until a ‘mirror finish’ is achieved In this way, we are able to make a conical head at home with specifications similar to those of studio-grade or top-of-the-range Hi-Fi heads
Transformer laminations are unsuitable for tape heads
heads -..........
make them yourself using transformer laminations
Do you know how thick transformer laminations are? And what is the thickness of the magnetic circuit in a four-track head (because I don’t suppose we want to make a mono one?) and, consequently, what is the thickness of the core laminations in such a head?
Furthermore, transformer laminations are quite difficult to machine due to their high silicon content – achieving a perfectly smooth finish under amateur conditions isn’t that straightforward. I’ll leave aside the idea of using resin to encapsulate the head – such a head would have a lifespan comparable to those from Kasprzak… So, at best, it will last for a few dozen hours before signs of wear on the face start to become visible. And one final point – the core in a universal, recording or reading head must have a slot perfectly perpendicular to the tape and filled with a bronze insert (approx. 4 micrometres thick, if I recall correctly).
In my humble opinion, given the above requirements, this idea is about as good as looking for a head from an M2405… although no… that head from Kasprzak at least had the advantage that you didn’t have to go to great lengths to get hold of it. That doesn’t mean, of course, that such a head is the eighth wonder of the world today – as I’ve written (and I say this from my experience as a repair technician), it’s a very…run-of-the-mill cylinder head; what’s more, there are significant variations between individual units, and in most cases 70 per cent of them have serious manufacturing defects. For example, incomplete filling – the face has a ‘crater’ where an air ‘bubble’ that remained inside during casting has been ground away. I shan’t even mention the crooked slots or the delaminated core plates. And what about Western-made heads? You’d be lucky to get one in good condition for a reasonable price...
make them yourself using sheet metal from speaker transformers on a permalloy core – I’m sure everyone has a burnt-out speaker transformer like that lying in a drawer, salvaged from the wreckage of some valve radio or valve television. You can make several dozen pickups from the core of a transformer like that.
Oh... I wouldn’t attempt that. I know it’s possible, and there have been descriptions of DIY heads in books for hobbyists, but it’s a hell of a lot of work.
or
make them yourself using sheet metal from speaker transformers on a permalloy core – I’m sure everyone has a burnt-out speaker transformer like that lying around in a drawer, salvaged from the wreckage of some valve radio or valve television.
Probably only in some high-end designs. I’ve never come across speaker transformers with such cores.
They are often sold at auctions in the former Soviet Union, and not for very much. From 10 euros each.
Leaving aside any aversion to that part of the world, there’s nothing to stop you from finding them in Poland too. It’s just that they’ll still be second-rate heads, and what’s more, not necessarily of good physical quality.
andreyatakum wrote:
Have you tried winding anything with 0.1 mm wire?
Contrary to appearances, a 0.05 wire diameter is quite manageable – I’ve wound pulse transformers with exactly that wire myself – by hand. And 0.1 is thick and strong enough that you can even use a drill to drive the coil. So winding the coil isn’t the biggest problem. The problem, if I may put it this way, is the mechanical part of the head – specifically, the precision of manufacture and dimensional accuracy. I won’t even mention grinding, because back in the 1990s there were already specialists who used this method to make new heads from old or worn-out ones and sold them as ‘brand new, never-used items’.
Do you know how thick the transformer sheet metal is? And what is the thickness of the magnetic circuit in a four-track head (because I don’t suppose we want to make a mono one, do we?)
The sheet metal from a speaker transformer is 0.3 mm thick
The thickness of the magnetic circuit is 1 mm per channel – for stereo, you need two magnetic circuits separated by a shield.
andreyatakum wrote:
398216 Usunięty wrote:
In my humble opinion, given the above requirements, this idea is about as good as looking for a head from an M2405.
They are often sold on auction sites in the former Soviet Union, and not very expensively. From 10 euros each.
Add to that VAT, customs duty and the cost of postage when importing into Poland.
andreyatakum wrote:
CHOPIN66 wrote:
a read head with coils of 4x1800 turns each, wound with 0.05 mm wire,
Have you tried winding anything with 0.1 mm wire?
Yes, I tried it once – I needed to wind a choke on a resistor for a super-reactive receiver. By the way, I know someone who has a winding machine.
andreyatakum wrote:
CHOPIN66 wrote:
on a permaloy core
Permalloy is very soft. A head like that won’t last long. In the later days of this technology, Sendast or even ferrite was used.
Just try and find Sendast anywhere in Poland. If you type the name into Google, you’ll only find bearings or seals
Permalloy in read/write heads, due to its magnetic properties, effectively suppresses eddy currents. So, in theory, the head would allow for read and write speeds of up to 22 kHz. A permalloy head will, however, show signs of wear after just 500 to 1,000 hours of continuous playback.
Opting for ferrites isn’t actually a bad idea either – it gets interesting because it turns out that the core material for an erasing head can be sourced from anti-interference filters fitted to USB cables – NiZn ferrites are used there, whilst the material for the write and read heads can be sourced from the cores of switching transformers used in the switching power supplies of computers and televisions – MnZn ferrites were used there. This means you don’t have to buy it.
fitted to USB cables – NiZn ferrites are used there
I suspect those ferrites are designed for high frequencies, as amateur radio operators often use them in the cores of amplifier output stage transformers and in antenna baluns. This means the magnetic permeability does not exceed u100.
CHOPIN66 wrote:
and for the read/write head, you can use the cores from switching transformers found in the switching power supplies of computers and televisions – MnZn ferrites were used there
Is it possible to build a high-end reel-to-reel tape recorder yourself? A beginner’s experience.
The answer is: you cannot build one. You can, however, improve the electronics of an existing tape recorder.
I once had ambitions to design such a tape recorder. As electronics students, we thought a tape recorder was just about the electronics. We believed that the decision to purchase a licence from Grundig was a mistake.
Today, I believe that the mechanics and technology are a much bigger problem than the electronics. The heads themselves are a major technological challenge.
Mechanics – I based my design on the “BG20 ‘Szmaragt’” tape recorder. This is what that very old design looked like.
I recently built a head amplifier using modern op-amps. Will such an amplifier be better than one based on the MPSA-18?
Added after 12 [minutes]:
CHOPIN66 wrote:
and for the write and read heads, you can use the cores from switching transformers
andreyatakum wrote:
That fits
Obtaining the material is not enough to make the head
The heads can only be manufactured by a factory with modern machinery. For example, precision grinding machines are required to grind the slot.
Is it possible to build a high-end reel-to-reel tape recorder yourself? A beginner’s experience.
The answer is: no, you cannot build one. You can, however, improve the electronics of an existing tape recorder.
I once had ambitions to design such a tape recorder. As electronics students, we thought a tape recorder was just about the electronics. We believed that the decision to purchase a licence from Grundig was a mistake.
Today, I believe that the mechanics and technology are a much bigger problem than the electronics. The heads themselves are a major technological challenge.
Mechanics – I based my design on the “BG20 ‘Szmaragt’” tape recorder. This is what that very old model looked like.
I recently built a head amplifier using modern op-amps. Will such an amplifier be better than one based on the MPSA-18?
Added after 12 [minutes]:
CHOPIN66 wrote:
and for the write and read heads, you can use the cores from switching transformers
andreyatakum wrote:
That’s right
Obtaining the material is not enough to make the head
The heads can only be manufactured by a factory with modern machinery. For example, precision grinding machines are required to grind the slot.
It is better to build the amplifier using transistors – even modern op-amps generally have high input resistance, which can cause noise issues.
High-end tape recorders had mechanics simplified to the point of absurdity – because they were direct-drive designs with three motors – the Pioneer RT 707 had virtually no mechanical parts at all; the only mechanical elements were the tape tensioning mechanism, braking, a counter and that’s it – the rest consisted of three motors (AC, DC, BLDC – depending on the designer’s choice) and two solenoids. The Szmaradg, Melodia, ZK-12O, Arie, etc. must have had complex mechanical systems because everything was driven by a single AC motor powered directly from 220V, which ran at a constant speed; consequently, friction gearboxes and belt drives were necessary to reduce the motor’s rotational speed – which, at the same time, complicated the mechanics.
And isn’t that the very essence of DIY? Exactly – the trick isn’t in buying second-hand or new ZRK heads, but in making your own with better specifications.
You sand them down with 3000–5000-grit wet sandpaper. We’re not talking here about a cassette tape recorder head, but a reel-to-reel tape recorder head. There were no factories specialising in tape heads, as was the case with cassette tape recorders – the factory supplied only the raw core, whilst the rest – sanding, winding the coils and encapsulation – was carried out by the tape recorder manufacturer. That is why there was no single standard for head mounting spacing; each manufacturer had its own. That is why, for example, the Chinese do not manufacture heads for reel-to-reel tape recorders – when their economy was opening up to the free market, reel-to-reel tape recorders were already in decline in favour of cassette recorders.
By using a microcontroller and a tension sensor, mechanical braking can be eliminated altogether – braking can be achieved using motors.
It’s better to build an amplifier using transistors – even modern op-amps generally have high input resistance, which can cause noise issues.
There is such a thing as a compromise – a transistor-based first stage followed by an OP-amp.
CHOPIN66 wrote:
High-end tape recorders had mechanics simplified to the point of absurdity
Because, over time, the most optimal design was developed. Unfortunately, such things take time to understand – what’s important and what isn’t. For example, a common problem with early designs was relying on power transmission via wheels with a rubber ‘rim’ – over time, it became apparent that, despite its undoubted advantages, rubber does not last forever – it lost its properties and slippage problems began to arise. Nowadays, rubber as such is used only where it is absolutely essential – to press the belt against the drive roller, for belts transferring power from the motor to the flywheels, and so on. Other parts of the drive train rely on gear wheels, which have proved to be far superior. Admittedly, the aforementioned Grundig made a ‘mistake’ by failing to take into account the ageing of the soft polymers used in the sprocket – which would have been the noisiest component had it been made of a hard polymer – but that was an exception.
CHOPIN66 wrote:
Exactly – the trick isn’t in buying second-hand or new ZRK heads, but in making your own with better specifications.
You sand them down with 3000–5000-grit wet sandpaper
And you cut out the shapes with scissors with the necessary precision... Oh, I almost forgot – you run the metal sheets you’ve previously obtained from a transformer through a sheet metal shearing machine to make them suitably thinner... Let’s not get carried away – theory is one thing, but eventually you’ll come up against certain difficulties that you can’t just sort out on the fly. It’s best to think your idea through properly first, and only then insist on its feasibility.
Exactly – the trick isn’t in buying second-hand or new ZRK heads, but in making your own with better specifications.
CHOPIN66 wrote:
You sand them down using 3000–5000-grit wet sandpaper.
You sand ‘by eye’ – without a precision sander.
Do you really think you can do that?
Whilst maintaining the flatness of the head’s front face. Do you at least have a tool for measuring the width of the gap?
CHOPIN66 wrote:
There were no cylinder head factories, as was the case with cassette recorders – the factory only supplied the raw core.
cyrus2 wrote:
Head assemblies can only be manufactured by a factory with modern machinery.
CHOPIN66 wrote:
the rest – grinding, winding the coils, and encapsulation – was carried out by the tape recorder manufacturer
The tape recorder manufacturer is also a factory.
A factory equipped with machinery and skilled workers.
DIY tape recorder works, but demands extensive iteration
Building a tape recorder – I’ve given it a go, and not just once. I think this is the third iteration of the idea (YouTube links below). It started as a studio-scale project, and I gradually scaled it down. I made the project easier for myself by using ready-made, recycled parts. What was the result? A working tape recorder. Would I make another one? Definitely yes. Would I recommend other people do this? I’d definitely advise against it. The number of swear words that were uttered during the build, and the amount of stress involved, was enormous. The result – needs work (which is how the next version will come about). No, I learnt a great deal from this project.
To answer the question in the title: It’s possible, but what’s the point?
With the first cylinder head bridge from the ZK140T, a three-head version with a belt that can be rewound by hand:
A stepper motor and any gearbox, even a planetary one, should produce the desired effect
A stepper motor jerks even in half-step mode. Add to that the gearbox and backlash in the gears – it’s not a good idea. On top of all that, there’s the problem (albeit not as significant) of filtering out debris from the motor’s power supply. It is for good reason that a series motor or an asynchronous squirrel-cage motor (with short-circuited rotor windings) is used to achieve perfectly smooth tape transport. Revox, Studer and other tape recorders (including even some vintage models – such as the Melodia) used three-phase* asynchronous motors for tape transport – the heavy rotor acted simultaneously as both the drive and the flywheel, resulting in exceptionally smooth tape transport without jerks or fluctuations in rotational speed – moreover, the mains frequency is now much more stable than it was 50 years ago.
* Of course, the motor itself was powered by a three-phase supply, even though the tape recorder was connected to a 220 V mains supply (as was still the case at the time). The phase shift was provided by a capacitor (for the second phase) and the inductance of the winding in one section of the motor (the third phase). The same method was even used in the aforementioned products from Studer and its slightly simpler counterpart, Revox.
Interestingly, in the licensed ‘Marcin’ model, the tape drive motor was also three-phase, but powered by a three-phase generator (to allow for speed correction) built using transistors... As the motor did not need to be particularly powerful (after all, the Marcin is a cassette player), this solution proved effective – a flywheel was added solely to improve rotational stability (as the rotor’s mass was relatively small and the tape recorder aspired to a higher class).
Rejecting stepper drives for smooth tape transport
A stepper motor and any gearbox, even a planetary one, should produce the desired effect
A stepper motor jerks even in half-step mode. Add to that the gearbox and backlash in the gears – it’s not a good idea. On top of all that, there’s the problem (albeit not such a big one) of filtering out debris from the motor’s power supply. It is for good reason that a series motor or an asynchronous squirrel-cage motor (with compact rotor windings) is used to achieve perfectly smooth tape transport. Revox, Studer and other tape recorders (including even some vintage models – such as the Melodia) used three-phase* asynchronous motors for tape transport – the heavy rotor acted simultaneously as both the drive and the flywheel, resulting in exceptionally smooth tape transport without jerks or fluctuations in rotational speed – moreover, the mains frequency is now much more stable than it was 50 years ago.
* Of course, the motor itself was powered by a three-phase supply, even though the tape recorder was connected to a 220 V mains supply (as was still the case at the time). The phase shift was provided by a capacitor (for the second phase) and the inductance of the winding in one section of the motor (the third phase). The same method was even used in the aforementioned Studer products and those of its slightly simpler counterpart, Revox.
Interestingly, in the licensed ‘Marcin’ model, the tape drive motor was also three-phase, but powered by a three-phase generator (to allow for speed correction) built using transistors... As the motor did not need to be particularly powerful (after all, the Marcin is a cassette player), this solution proved effective – a flywheel was added solely to increase rotational stability (as the rotor’s mass was relatively small and the tape recorder aspired to a higher class).
And what if such a stepper motor were powered by sine-wave pulses? Then we’d maintain smooth motion because the waveform rises and falls smoothly, and we’d eliminate the jerks that occur with PWM and square-wave signals.
Tape recorders used both three-phase and induction motors
or an asynchronous squirrel-cage motor (with short-circuited rotor windings). Revox, Studer and other tape recorders (including even some vintage models – such as the Melodia) used three-phase* asynchronous motors
In the case of the Melodia, an SS-60 motor was used, which was a synchronous motor with asynchronous start-up. Furthermore, it had two rated speeds. You’re actually right – it was a three-phase motor, powered from a single phase. At least, that’s what the source I have also states.
However, in most tape recorders I’ve seen, asynchronous motors with a short-circuited auxiliary phase were indeed used. The same applies to older-generation record players (e.g. the entire Bambino series).
ADVERTISEMENT
Clarifying asynchronous versus permanent-magnet synchronous motors
However, in most of the tape recorders I have seen, asynchronous motors with a short-circuited auxiliary phase were indeed used.
I was taught that a synchronous motor isn’t necessarily one with a squirrel-cage rotor, because in such a motor (even if it has a short-circuited stator winding) there is always slip – that is, a lack of perfect synchronisation of rotation under load. But OK, maybe I’m wrong. The only synchronous motors (or at least those that deserve the name based on their design) are motors with a permanent magnet rotor – such as those used, for example, in most low-speed AC motors.
In such a motor, the rotor’s magnetic field is much more strongly coupled with the stator’s magnetic field, which ensures much greater speed stability under varying loads.
The only synchronous motors (or at least those that deserve the name on account of their design) are motors with permanent-magnet rotors – such as those used, for example, in most low-speed AC motors
The rotor in a synchronous motor does not necessarily have to be a magnet; it can be made of special steel. This synchronous motor from the Melodia tape recorder was quite a curiosity, as the rotor was located on the outside of the stator. More on this topic here:
Asynchronous motors, on the other hand, are designed to operate with a certain slip relative to the frequency of the supply current in order to generate torque. However, if they have so-called ‘hard’ characteristics, this slip remains relatively constant as a function of load. This is why they worked so well in the tape-drive mechanisms of cassette recorders.
That is why they worked so well in the tape-drive mechanisms of cassette recorders.
In my opinion, they were simply a cheaper alternative to a motor. But – of course – that’s just my opinion; however, given the design of top-of-the-range tape recorders, there must be some truth in it, surely?
but I understand perfectly.
