[AudioI] Guitar Preamp for Isolated(Pro) soundcard

Sun Mar 6 19:13:47 UTC 2022

Well, if it works, it works. I’m not any kind of expert in analog design, just trying to learn. In fact, it seems like a very good design in terms of designing a preamp board.

Can I ask about the output circuits in your designs? I think I understand how to deal with the filtering requirement on the input circuit in the codec data sheet, but the data sheet  just gives an example circuit for the output and some vague wording about compensating for the DAC without actual specs. That circuit looks a bit weird in terms of how unsymmetrical the filtering is on the differential channels. Do you just design a DC coupled version of that, or is there some underlying principle you use that gives you more freedom of design?

Thanks again,
Ted

> On Mar 5, 2022, at 7:47 PM, Matt Flax <matt at audioinjector.net> wrote:
> 
> Yes the preamp's GND power floats with Vcom. It is always at digital zero (roughly).
> 
> The power lines on the preamplifier and the audio codec are isolated for that reason close to no power should be required by the codec to maintain Vcom.
> 
> As the power lines on the preamplifier and the audio codec are isolated, there should be minimal problems with ground loops. However I would like to hear if there are such problems.
> 
> Indeed the entire analogue audio subsystem has its power lines isolated, the only reference between the Pi and the audio codec are their ground pins. I guess you could call that floating. The original Isolated sound card was entirely floating. There wasn't even a GND line connected between the two systems.
> 
> How does that sound to you ?
> On 5/3/22 05:49, Ted Rippert wrote:
>> Matt,
>> 
>> Thanks for the info. I get that Vph is phantom power, nice option. 
>> 
>> I’m not sure I understand about the GND being at Vcom. Does the ground of the preamp card float 2.5V above the ground of the Pi? Seems like this would be a problem when you connect a shielded audio cable to some other equipment as you could get some large currents through the shield. Also, there is a limit of 1 uA on the DC source/sink current of Vcom. So, I’m sure I’m misunderstanding something here. Maybe you’re floating the whole analog section of the Codec chip so it is not         referenced to the Pi ground at all?
>> 
>> -Ted
>> 
>>> On Mar 4, 2022, at 4:07 AM, Matt Flax <matt at audioinjector.net <mailto:matt at audioinjector.net>> wrote:
>>> 
>>> Yes that is right there is Vcom, but don't use Vph for that.
>>> The ground of the preamplifier connector is signal ground (Vcom). To be more specific, the GND lines in this schematic are regarded as a zero digital signal out of the DAC and into the ADC : https://github.com/Audio-Injector/AudioInjector-Analogue-Breakout-Standard/blob/master/AudioInjectorAnaBreakStdv1.pdf <https://github.com/Audio-Injector/AudioInjector-Analogue-Breakout-Standard/blob/master/AudioInjectorAnaBreakStdv1.pdf>
>>> This means, you don't need to worry too much about ground levels shifting along the chain. In the end is makes preamplifier design more trivial, as long as the lines connected to the codec are balanced (differential).
>>> 
>>> Vph is a 48V line for phantom power - it can only supply very very low current, but a large voltage.
>>> 
>>> The AI-ABS has 4 voltages on it :
>>> 
>>> +-12V (which are actually larger then 12V so you have preamp design headroom for opamps). GND which is balanced signal ground. There is the Vph of around 48V. There is also the dirty Pi 5v line.
>>> Regarding balanced codec lines, if the input or output connections are single ended, then you need to ensure that the codec connections are differential. For the input opamps, you can use an inverting unity gain opamp or a negating voltage follower to create the inverted input. However there may be more sophisticated methods for achieving this with instrument inputs - I am not too sure. I would imagine there are some sophisticated multi opamp instrument input preamplifier circuits out there.
>>> Matt
>>> 
>>> p.s. I like KiCaD because anyone in the world can use it and that means our circuits are shareable without concern of the economic circumstances of the other party.
>>> On 4/3/22 21:23, Ted Rippert wrote:
>>>> Thanks Matt. That looks like a useful filter. I guess that is what those two big caps are for on the RCA preamp.
>>>> 
>>>> I was wondering about the power pins on the Analog Breakout schematic. There is +12V, -12V, GND, and Vph. What is Vph? I was hoping it might be Vcom (Common Mode Voltage) from the CS4272.
>>>> 
>>>> I don’t use KiCAD, but I should probably try it. It looks like it does pretty much everything. 
>>>> 
>>>>> On Mar 4, 2022, at 2:25 AM, Matt Flax <matt at audioinjector.net <mailto:matt at audioinjector.net>> wrote:
>>>>> 
>>>>> Some good information in this thread.
>>>>> 
>>>>> For the power supply (from the Pro), you will want a simple low resistance filter on both the +- rails. For those you will want to check out "capacitance multiplier" filters which are typically a NPN (or PNP) transistor, a resistor and a large electrolytic capacitor in parallel with a 100 nF ceramic. The two capacitor types in parallel are meant to approach a more ideal capacitor. The circuit is very simple and can be seen in this mail and at the top of this link for the positive rail : https://electrical.codidact.com/posts/279580 <https://electrical.codidact.com/posts/279580>
>>>>> 
>>>>> 
>>>>> Some people add extra resistance at the base or other things. You can probably find lots of references online.
>>>>> For the negative rail, switch the type of transistor (PNP) and mirror the circuit about signal ground.
>>>>> 
>>>>> Oh and the supply voltages are > +-12V so make sure the opamps you choose can do the necessary +-20 volts or so. This means that your opamps will have plenty of headroom for the signals you are planning to put through them.
>>>>> Someone want to start a KiCad repo somewhere ?
>>>>> 
>>>>> Matt
>>>>> 
>>>>> On 4/3/22 20:20, Ted Rippert via People wrote:
>>>>>> Thanks. I have seen this circuit before, but forgot about it. It could make a nice input circuit if you change the 10k input resistor for a 1Meg and set the gain with an appropriate value of feedback resistor instead of the 10k there. The codec has a common mode voltage output that could be used to set the bias point. Should also be possible to add a small Capacitor in the feedback loop to get the filtering that the data sheet asks for.
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>>> On Mar 3, 2022, at 6:31 PM, Joseph DiVerdi <joseph.diverdi at colostate.edu <mailto:joseph.diverdi at colostate.edu>> wrote:
>>>>>>> 
>>>>>>> Here's an example of a circuit that can accomplish what you want.
>>>>>>> 
>>>>>>> <image0.png>
>>>>>>> If IC1and IC2A are eliminated you have a robust single to differential converter. Easy resistor values. If you use bipolar power you can take the plus terminal of IC3B to ground in the outputs will be centered around 0V. It's DC coupled also. You can use the op amps of your choice. I'm partial to the Analog Devices ones because extremely high quality  along with good prices. But like I say you can use whatever ones you like.
>>>>>>> 
>>>>>>> jadv
>>>>>>> 
>>>>>>>> On Mar 3, 2022, at 5:55 PM, Ted Rippert via People <people at lists.audioinjector.net <mailto:people at lists.audioinjector.net>> wrote:
>>>>>>>> 
>>>>>>>> Lots of good information so far, but I do want to get back to the main purpose of this thread. Making a preamp specifically for the Pro sound card.
>>>>>>>> 
>>>>>>>> Since the Pro already has +-12VDC on it, I’ll assume that the power supply is taken care of.
>>>>>>>> 
>>>>>>>> One thing I would like to get more information on is a good opamp circuit to interface a passive pickup to the differential inputs available on the Pro card header. The CS4272 data sheet gives an example of a circuit to use for the analog input, but it is differential to differential, and AC coupled. There is also a recommended output circuit that is AC coupled. The specs on the Pro card and the existing Preamp boards for it say that they are DC coupled. I like the idea of DC coupling, especially for a Bass, but what are the technical details. The data sheet does state that you need to filter out 6.144 MHz and it’s harmonics, which is fairly straight forward with an opamp. There are various differential to single ended and vice-versa circuits documented on the web, such as:
>>>>>>>> 
>>>>>>>> https://sound-au.com/articles/balanced-io.htm#s1 <https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsound-au.com%2Farticles%2Fbalanced-io.htm%23s1&data=04%7C01%7Cjoseph.diverdi%40colostate.edu%7C05245f5d39f54d5c4b5508d9fd799a0c%7Cafb58802ff7a4bb1ab21367ff2ecfc8b%7C0%7C0%7C637819521003196260%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=usFmKT67YoY0g9GSXaxBAofIyozEBKarvWE0TGM%2BIpQ%3D&reserved=0>
>>>>>>>> http://www.douglas-self.com/ampins/balanced/balanced.htm <https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.douglas-self.com%2Fampins%2Fbalanced%2Fbalanced.htm&data=04%7C01%7Cjoseph.diverdi%40colostate.edu%7C05245f5d39f54d5c4b5508d9fd799a0c%7Cafb58802ff7a4bb1ab21367ff2ecfc8b%7C0%7C0%7C637819521003196260%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=mO8RKdTmyC9ZefmASyxbfgIzOWe5mierhAEgHIdpYoU%3D&reserved=0>
>>>>>>>> https://sound-au.com/project87.htm <https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsound-au.com%2Fproject87.htm&data=04%7C01%7Cjoseph.diverdi%40colostate.edu%7C05245f5d39f54d5c4b5508d9fd799a0c%7Cafb58802ff7a4bb1ab21367ff2ecfc8b%7C0%7C0%7C637819521003196260%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=KL9Nt8Fhlt9NB3ivuYhjyRGL2BHb8fRvq4tLn%2Ba6gpI%3D&reserved=0>
>>>>>>>> 
>>>>>>>> 
>>>>>>>> Does anyone have experience into how to modify such circuits for a high impedance input while maintaining the needed filtering and possibly adding a gain control on the input? I know the theory on how to do such things, but there are a lot of possible options and the devils are always in the details. Also, are the pins on the Pro card header tied directly to the Codec chip, or is there some circuitry in between? I’m going to order a Pro board and one of the preamp boards for general use and to see what I can find out about the circuits and layout. Just hoping to get a bit of a head start here.
>>>>>>>> 
>>>>>>>> Thanks, and I don’t mean to stop any of the other conversations in this thread. They are very useful to me and hopefully others. Just want to also try to get some info on the main subject.
>>>>>>>> -Ted
>>>>>>>> 
>>>>>>>>> On Mar 2, 2022, at 2:41 PM, Ted Rippert <ted.rippert at gmail.com <mailto:ted.rippert at gmail.com>> wrote:
>>>>>>>>> 
>>>>>>>>> Hello,
>>>>>>>>> I’ve seen several threads on this list relating to using Guitar or Bass with an Audio Injector card. I’ve been doing that for years using a rack mount mixer with a couple High-Impedance inputs interfacing into an Audio Injector Stereo card. Works very well, and I can get round trip latencies on the Pi as low as 2.1 ms running Guitarix amp sim if I use some CPU isolation techniques. 
>>>>>>>>> 
>>>>>>>>> I’ve been playing around with making a small preamp/buffer that runs on the 5V from the Pi, and found just what others have, that it’s too noisy. So I worked up a setup with a dual supply converter that then goes into a couple linear regulators. It does work, but it’s a mess using just perf board and through hole parts. Also, the Zero sound card I used has the same usable but high noise floor that the stereo card does.
>>>>>>>>> 
>>>>>>>>> So I want to try making a preamp for the Isolated card (really a Pro card from Amazon, but same interface). I would like to talk to Matt and any others with experience about the design of such a card, and any possibility of an “official” Audio Injector version being made. I just do perf board/through hole stuff, so I can only make one-offs myself.
>>>>>>>>> 
>>>>>>>>> To get things started, let me list some specs for such a card:
>>>>>>>>> 
>>>>>>>>> Input impedance: > 500kohm, 1 Mohm nominal
>>>>>>>>> Input voltage: ~ 1 VDC pk-pk, preferably with an adjustable gain and significant headroom
>>>>>>>>> Input connector: 1/4 inch TS socket
>>>>>>>>> Output impedance: low audio interface level, preferably less than 300 ohms
>>>>>>>>> Output voltage: at least +4dBu, preferably with adjustable gain, since the codec has an adjustable analog output volume, this may be in software
>>>>>>>>> Output connector: 1/4 inch socket, preferably a TRS type that will work as both a balanced and unbalanced output depending on what is connected to it
>>>>>>>>> 
>>>>>>>>> It could be as simple as a modified RCA preamp with different connectors and higher impedance input circuits, or you can think about adding things like a DIN based MIDI interface and a headphone amplifier output. All of this seems quite doable using the AI-ABS standard, but I need some more details about the specs of that, and about the practicalities of how to design opamp buffers to interface to the codec on the Pro board.
>>>>>>>>> 
>>>>>>>>> So, if anyone is interested, lets talk about it.
>>>>>>>>> 
>>>>>>>>> Thanks, 
>>>>>>>>> Ted
>>>>>>>> 
>>>>>>>> 
>>>>>>>> -- 
>>>>>>>> People mailing list
>>>>>>>> People at lists.audioinjector.net <mailto:People at lists.audioinjector.net>
>>>>>>>> https://lists.audioinjector.net/mailman/listinfo/people <https://lists.audioinjector.net/mailman/listinfo/people>
>>>>>> 
>>>>>> 
>>>>> -- 
>>>>> Checkout the community email list :
>>>>> https://lists.audioinjector.net/mailman/listinfo/people <https://lists.audioinjector.net/mailman/listinfo/people>
>>> -- 
>>> Checkout the community email list :
>>> https://lists.audioinjector.net/mailman/listinfo/people <https://lists.audioinjector.net/mailman/listinfo/people>
> -- 
> Checkout the community email list :
> https://lists.audioinjector.net/mailman/listinfo/people <https://lists.audioinjector.net/mailman/listinfo/people>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.audioinjector.net/pipermail/people/attachments/20220306/eb198625/attachment-0001.htm>