Test & Measurement interface for Soundcard

Thanks Tubelab,
Cool stuff, thank you for sharing your understanding and experience. I do not understand all that I know about the interaction between/among the parts of the SET amplifier and the loudspeaker output device. Thanks to Thiele and Small we know that speakers and their boxes have Q. Amplifiers have Zo. I have not seen anyone tie the amplifier and speaker together in the concept of the total Q of the system. What I have seen is the operating assumption that amplifier has a Zo of 0 ohms and the speaker including box/baffle has a Total Q of 0.71. SET’s do not have a Zo of zero.
Bass is good, flabby is not.
The Millett interface and the dedicated PC will get a workout at my house.
DT
All just for fun!
 
My experience as an RF engineer has offered me another solution. I use an RF spectrum analyzer to poke around inside every one of my tube amp designs. The little 5842 that I use in the Tubelab SE can easilly sing away at 400 MHz. Just touching the grid pins with the scope probe kills the oscillation, but it is plainly visible with the spectrum analyzer's input placed in the vicinity of the tube, without physical contact. It would be beyond the expectations of a casual amp builder to do this, but many of these oscillations can be detected with a radio or analog TV placed in close proximity to the amp.

I have a couple of RF Sniffers, one was made with the flying head of a hard disk drive (I haven't been able to find that one for a while), another is some wire-wrap around a ferrite (like an a.m. antenna). I have seen one made with an Analog Devices detector/log converter. If you have a spectrum analyzer (I have an HP3577) you can poke around in the amplifier and see what's really going on. It's interesting to look at the switching noise on semiconductor diodes -- and see how these are attenuated.
 
I have a couple of RF Sniffers, one was made with the flying head of a hard disk drive

I have a couple of them too. The one I use most is simply a piece of coax cable with a BNC on one end and the other end has the braid stripped back to expose about 2 inches of the center conductor. The end is covered with heat shrink to avoid fireworks when poking around in live equipment. The other has a coil with a ferrite core attached to the business end. I have no idea what the coil came out of.

I have an old HP141T mainframe with the 0 to 1.25 GHz RF plug in. I made it out of the remains of two millitary scrap units that I got cheap back in the 1970's when these units sold for about $3500 used. They can often be found for $200 to $500 now. The analog storage CRT on these units are rare so make sure that you verify functionality before buying a used unit. Don't expect the storage function to work too good, though. It sucked even when these things were brand new.

It is amusing to see just how much RF garbage comes spewing out of some common electronics. The worst offender is the PC. And guess what powers that sound card. My old PC audio analysis setup always had a few spikes in the 18 to 25 KHz region. These came from the CRT based monitor. I always had to turn the monitor off when saving plots for posting to my web site.

Cool stuff, thank you for sharing your understanding and experience.

I have a bunch of material of this sort that will be posted to my web site when I finish rebuilding it all.

I do not understand all that I know about the interaction between/among the parts of the SET amplifier and the loudspeaker output device
.

That particular interface is a very complex one. The most often overlooked phenomena are the fact that the speaker drivers are also microphone devices, and they have a DC resistance. The drivers can each produce a voltage in response to air pressure on their cones, and in response to mechanical vibration from the cabinet and other drivers inside the cabinet. Many users spend tons of time and money trying to achieve a very high damping factor (low output impedance) and often look for speaker wire with an extremely low resistance. It is important to remember that this is all effectively in series with the DC resistance of the drivers, which can be several ohms. A low output impedance is however important to prevent any EMF (unwanted microphone voltage) generated in the speaker system from making it back into the amplifier.
 
Pete Millett,
CQ CQ are you monitoring Pete?
I have a question about SW4, the Grayhill 2p 6t rotary switch. I have the Millett interface powered up. SW4 looks to be a quality part selection. It comes with two small pins that insert into the switch body to limit the number of active poles. I want to only take this thing apart once to set the pins.
This is where I start making assumptions. The switch will have 4 active positions. The first, left most position will have “mV” in the lower right of the display. The next three positions as the switch is turned clockwise are the 2V,20V & 200V positions these 3 positions will all have “V” in the lower right of the display. Am I on the correct page?
BTW: The enclosure and part pieces including display all fit together well. Good job.
DT
All just for fun!
 
It comes with two small pins that insert into the switch body to limit the number of active poles. I want to only take this thing apart once to set the pins. This is where I start making assumptions. The switch will have 4 active positions.

Oh yeah, I didn't really mention that, did I?

Yes, the switch stops should be set for 4 positions. Position 1-4 are used, so you can follow the instructions that came with the switch as to how to put those little tiny pins in.

If I remember right, you make sure the flat on the shaft is opposite the number "2" on the back. Then you insert pins just to the left of number "1", and to the right of "4", as viewed from the shaft end. Then you stick the shiny label thing on the front of the switch body to hold the pins in.

I always make sure the pins are right by checking the switch rotation before sticking the label on.

Be careful, those tiny pins are hard to hold on to, and they disappear from the face of the earth if you drop them! I have sucessfully substituted a short piece of piano wire or other steel wire...

Pete
 
Hello,
For temporary use I tested the switch with pieces of wire that were clipped from one of the ¼ watt resistors. That wire is most likely to soft to last but is near the correct size.
My guess at the size is closer to 0.010 inches perhaps a red hair more. Tubelab in your world a piece clipped from a guitar string may work.
Hint, look in the Dgikey box.
DT
Just for fun!
 
All the knobs I have use a set screw that is 180 degrees off from the indicator mark. If I set the knob with the set screw on the flat of the switch it appears that the ranges are off. I removed the knob, adjusted for what looked like the 200mV range and reset the knob to point to the 200mV range.

Is this correct? It appears to me that the ranges are set by the layout and therefore, this must be correct.

No?
 
I swapped the 74HC86 in case it was bad and got the same results so the knob position appears correct. The unit powers up and runs with the scales shifting the decimal point as expected.

However I get about 2X the value I expect at 1KHz 190mV RMS in. I'm using the B&K 2534 oscilloscope in RMS measurement mode and comparing it with my Fluke 8020B to verify the RMS value and both agree within 2% so I'm confident with the input value.

When I went to install the resistors I bought locally in packs of 4, and I measured each resistor and matched them when they were in the diff-amp circuit. When installing them in most cases they were within .05%. I used multiple resistors in parallel to set the value to the value specified (R17).

The only carbon resistor is R15 (34.7K vs 34.8K).

The only substitutions made were all 150uF caps vs specified 330uF, and 1.5uF tant for C21 (1uf) and C12 & C10 (2.2uF). I looked at he AD538 data sheet and it looked like I could change these values without effecting operation.

I also added some 1nF bypass caps in parallel with the 100nF ones (on the back of the board) as experience has shown in the past that the combination of 100nF plus 1nF gives superior noise suppression and better decoupling.

With 960mVRMS in I have 440mV RMS at pin 5 of U7. This looks like it is twice what I expect (960mv/2V * 500mv = 240mV RMS expected) , so I must have something wrong in the input stages.

I guess I need to go back and re-check the resistor values.
 
However I get about 2X the value I expect at 1KHz 190mV RMS in. I'm using the B&K 2534 oscilloscope in RMS measurement mode and comparing it with my Fluke 8020B to verify the RMS value and both agree within 2% so I'm confident with the input value.

Nothing jumps out as obvious.

I would test with all three meters connected at the same time, just to make sure something isn't loading the generator. Safest to test with one side grounded to make sure there's not some common-mode issue.

On the 200mV range, with a 200mV input, there should be 500mV at the output of the INA134 (U7 pin 6). If there isn't, either a resistor is wrong, or a solder joint is bad, or something really odd is oing on...

Pete
 
Administrator
Joined 2004
Paid Member
Got my M-Audio 24192 sound card today - installed it in the successfully repaired eBay mule I mentioned in an earlier post. (Antex Aria case with newly installed psu, Athlon 64 XP3000, Abit KV-80 Mobo, Hitachi 160GB HD) Everything is working quite well, and the machine is very fast. Amazing how quickly XP boots with nothing else but the sound card and audio software installed. (About 6 seconds!) Since the machine is not going to be connected to my internal wireless network I don't even need virus software.. Acquisitions if I need to print them will be converted to pdf using a pdf converter and schlepped via thumbdrive to a networked machine (linux box) for printing.

Now I just need to get a small LCD monitor and build Pete's box.. I got a tiny keyboard with integrated touchpad to keep things simple.

Audio Tester 2.2d is a bit buggy, although much better than version c in this regard. I am not using asio because audiotester crashes pretty consistently when I try to use it. (This is true on two other machines as well.) I have locked the sample rate at 192K and audiotester is configured for 24 bits and 192kHz with the WDM drivers which works well. I have learned the hard way that after making a configuration change to immediately exit and restart the program - this results in pretty stable behavior.
 
Last edited:
Thanks all.

Resistors check as expected in the circuit (Low vs specified values but not abnormally so). I don't see anything out of the ordinary when rotating the switch through 360 degrees.

However U4 gets very warm to the touch (130F+? guestimate). I think I may have zotted the ic when installing it without a ground strap. Very dry cold weather that day. Bad for ESD.

I'll order a couple spare OPA2134s and a spare INA134 as well and see if that makes a difference.