I decided to share a problem that I encountered when I became the owner of the Behringer UMC202HD rev.1.12 card, and describe the way I solved it.
Maybe some of you will need it.
First things first. Back in the day, looking for a card / interface for audio measurement, and on a budget, I decided to purchase the inexpensive Behringer UMC202HD. Then I took it for granted that since it is compatible with the 192kHz / 24 bit technology, it must have satisfactory parameters. Unfortunately, the reality turned out to be completely different. The very first test of the card with RightMark Audio Analyzer showed its weakness. The test results showed that it was not suitable for the purpose for which I purchased it. Amazed by this fact, browsing through various entries about it on the web, I found, among other things, a certain comparative test of several different cards, which unfortunately confirmed my disappointment with this Behringer: https://www.infomusic.pl/test/46209,test-interkietasy-audio-behringer-focusrite-presonus-tascam It must be admitted, as it seems to have been mentioned in those tests, that with this card, each minimal change in the position of the input regulators results in a significant change in the obtained results. Here are my tests of this card, compared to the much older, cheaper and simpler UCA222 with PCM2902 on board.
At that time, I even thought of withdrawing from the purchase contract and returning it to the seller. However, I decided that I would not give up and I would find and eliminate the cause of the weakness of this card. After a short "fun" I came to the conclusion that its biggest weakness are the input amplifiers (MIDAS, which was mentioned in advertisements as one of the advantages of this card). As an interesting fact, the tests reveal that the MIDAS input amplifiers generate a large level of the second harmonic (generally even). Perhaps the creators wanted to give them a tube sound. To find out about the aforementioned weakness of these amplifiers, it is enough to connect the card to the USB connector of the computer, connect headphones to it, but do not connect anything to its inputs. Then turn on the Monitor functions and increase the settings of the input controls. You can hear how "beautiful" it generates large noises, which are immediately silenced after turning off the Monitor function. There was one conclusion - you have to omit the input amplifiers completely. (To simplify the process of working out the card, I sent an e-mail to the Polish distributor with a request to provide a part of the schematic containing the input amplifiers, unfortunately it remained unanswered.) So I rolled the card (interestingly, the card was built on systems with quite good parameters), looked, "figured out" the layout a bit, thought and figured it out. I found a very simple way to bypass the entire input amplifiers and feed the signal directly from the inputs to the signal symmetrization circuits (implemented on IC8-AD8694) for the IC7-CS4272 converter. Such a solution could be called a direct line input. It had the effect I needed. In this way, I obtained a card with parameters suitable for using it to perform audio measurements. In order not to lose the factory capabilities of the card, which may be useful for other purposes, I designed a simple module that allows you to switch the card's functionality. Of course, between the factory and the new modified, with a shortened entry track.
Below I will describe how to modify the card, but first a few important notes on this topic. - Deciding to modify, you lose any warranty. - Modifications, you carry out on your own responsibility, you are responsible for various effects associated with it, failed activities, etc. - The modification works when you give the input an unbalanced signal with a jack plug (or an adapter of this type). - Switching on the modification, completely bypasses the input amplifiers, no manipulators, no input amplifiers. - Attention!!! When using the modified mode, giving too high input signal can damage IC8 - AD8694 !!!
Description of the modification Diagram of an additional module:
Placement of elements on the board:
Below in the PDF attachment with the mosaic of the tile. And this is an example of an assembled module.
List of necessary elements: SMD relay OMRON G6S-2F 5V (or equivalent) - 2 pcs, SMD capacitor 47uF 16V, preferably bipolar - 4 pcs, SMD capacitor 1206 100nF - 1 pc, SMD diode 1N4148 - 1 pc, Additionally, you will need: a single-section switch, goldpin pins (standard 2.54mm pitch), a piece of a thin two-core cable.
Modification steps: - Remove the following electrolytic capacitors from the card motherboard: C81, C85, C34, C35
- In place of the capacitors, solder long pins (Gold-pin)
- On the previously installed pins, mount and solder the assembled module.
- Connect the + 5V power supply to the module. I propose, as I did, from the bottom of the board to the pins of the C7 capacitor
- install the card function switch in the rear housing panel. The ON position enables modification-shortened input path. An example of a photo montage. The switch should be connected to the connector on the added module.
Rear view of the device after modification
Before trying to work, I suggest you check the correctness of all steps! When the switch on the back wall is turned on, the card has a new functionality (shortened input path), while when it is turned off, it is compatible with the original version.
Below, in the appendix, the full test result showing the differences, the green diagram in the appendix marked with the annotation V2, is the result with the attached modification. And here is a summary, showing the differences in the results of the test.
We've talked about this before. A great modification worth spending a few free hours because it translates into real value.
However, I wonder if every Behringer model has such a disadvantage. If so, it proves the producer very badly that having much better parameters of the oil within reach, the subject leaves such a "groats". I would like to believe that you were unlucky and your version of the device was some kind of unlucky batch, because this device does not cost PLN 3.50 as the USB card that I once measured and the noise level is similar (of course, the other parameters are also kicked up).
In my opinion, the chance to significantly improve the parameters and to protect the transmitter inputs was wasted, you had to add amplifiers with good parameters on the board or even replace them on the motherboard.
I am curious what operational amplifiers etc. are there
Integrated circuits (not all, let's say only from the audio path) present in this device - operational amplifiers are mainly quad AD8694 - 4 pcs (input amplifiers - IC9, signal symmetrization for ADC - IC8, filter and output buffer - IC11, headphone amplifier - IC13), but also NJM2122 - 2 pcs (it looks like they also work in amplifiers input), and 4580 - 1 pc (it looks like it works in the polarization system - artificial ground). - ADC / DAC chip is CS4272, - DSP and USB support - XMOS SK1814, - keys to the Monitor function - HEF4066.
(...) adding goldpins to it and using them simultaneously as a signal interceptor and a construction element
Considering all the aspects, what I wanted to do, how much space was available, etc., I came to the conclusion that this was the best solution.
I miss an additional control on the front panel that would remind the user of this fact with its illumination.
Or a limiter that would simply prevent damage
I thought about the control for a moment, but somehow I let it go. As for the simple limiter, I tried one option, but gave up because it did introduce some distortions. A second idea was born, but I haven't had time to test it yet. As if what, I will inform you.
does every Behringer model have such a disadvantage
Actually, this also makes me wonder. Therefore, in addition to the symbol, I also provided the version number of this card. I also hope that maybe other owners of such cards will give their opinion.
you had to add amplifiers with good parameters on the board or even replace them on the motherboard.
Well, in this case, it would rather be necessary to redesign at least a part (with input amplifiers) of the motherboard for a completely different one. Buddy, it's not a matter of the elements used, but how they are used (such elements also work in other cards that achieve better parameters than this one). I am not forcing anyone to duplicate this modernization, I am not even encouraging it. I have presented it so that those who encounter a similar problem have a simpler way to solve it. After a thorough analysis of my problem, I realized the nonsense of "digging" in the motherboard and I chose this shortcut of the card input path as the best solution. I believe that was the best solution here. But others have the right to try to solve it differently.
Coming back to the topic of the lack of securing the card entrance. When publishing a description of this modification, I had to include such a comment. However, I wonder how the original input amplifiers of this card are protected (I honestly did not go into it) and other similar cards. Sometimes you hear that someone burnt the card somewhere because he controlled it with too much signal.
A very interesting modification! I have the same card, but probably in the 16bit version, I use it with a measuring microphone to adjust / set the dsp in Helix / Match car amplifiers. Tomorrow I will check the series of my device, I wonder if in my application the improvement of parameters is so important and how to avoid giving too high a level? EDIT: I misled:
I have the UMC22 version. I need to check what's inside.
I think that minor corrections (but in this case too little) would not significantly improve the situation, or at least not as much as I needed.
Which operational amplifiers would you recommend, my friend? The ones that sit there at the factory do not have outstanding parameters, but it cannot be said that the general factory condition of this card is due to their weaknesses.
I did not work out the input amplifiers of this card in detail (because it was a waste of time), but I have the impression that the input sensitivity switches work in such a way that higher signals are simply suppressed to the microphone level, and then it is amplified and hence the high noise and distortion. If the gain was limited then, and not the input attenuation, it would be completely different.
Either way, my way of dealing with the weaknesses of this card is some alternative. I am not going to convince anyone of this approach, but for me it is very simple and enables the use of this card for the audio measurements I needed. Besides, with such measurements, it is even better if we shorten the audio path of this card.
This is why I am asking because it is difficult to find such a cheap audio interface, if it changed it a bit, it would be the most profitable Maybe he will come to me on Monday, so I will sit and see how they did it.
How about bypassing their entire mic preamp and putting preamplifiers on the SSM2019? I know it's a cost, but it's still minimal considering the price of audio interfaces.
SSM2019 have min. supply voltage - symmetrical 5V, so there will be a problem, because only unbalanced + 5V cards on board.
And let me remind you that in my solution:
I found a very simple way to bypass the entire input amplifiers and feed the signal directly from the inputs to the signal symmetrization circuits (implemented on IC8-AD8694) for the IC7-CS4272 converter. Such a solution could be called a direct line input.
So it completely bypasses their preamplifiers. And the balun is needed because the CS4272 "wants" a symmetrical signal at the input. But you can still try to change the original AD8694 for something better.
I think you'll show off your insights when you start to "play" with this card.
I use microphones quite often so the noise from this input is important to me. The fewer the better Symmetrical power supply is not a problem, now there are plenty of 5V converters> symmetrical 12V, e.g. in Google, LCT-156 designation.
Then we should get a nice interface for measurements as well as a nice one for recording sound from condenser microphones.