that1510 based mic preamp problems

Hi All,

I'm messing around with a THAT1510 preamp that is a "mash-up" of fig. 4 from the THAT1510 data sheet from +IN/-IN to R1/R2 with the rest of the circuit taken directly from Jung and Garcia's Analog Devices AN-242 app note.

The circuit is currently powered from 2 x 9V batteries giving approx +/-9V rails.

I'm pretty happy with the performance considering the construction methods, however after recording the preamp with a 150ohm resistor bridging the +IN/-IN and examining the spectrogram of the recording I've noticed that there is a distinct spike at 15626hz.

I'm sure that the construction is not helping here(!!) but I curious if I was doing something obviously wrong??

cheers
Paul

edit - misleading spectrogram removed, see later postings.

Are you near a television set?

TV has been switched off at the wall for the past few weeks. I _was_ recording near my laptop however. I'll try repeating the process away from other electronics, however I am recording using a sony MZ-RH1 HiMD recorder and it had crossed my mind that the interference was being caused by the recorder.

If it's interference being caused by nearby electronics, I'd guess the situation will improve with the pre-amp in a case with some decent shielding?

cheers
Paul

Actually that first spectrogram was from a file recorded preamp -> line in on my powerbook.

This one is preamp -> mz-rh1 line-in recorded a couple of rooms away from the nearest electronics:



The pre is running at 66dB gain and the RH1 line-in is adding somewhere between 13-15dB gain, so the unweighted EIN is around -128dB. The laptop was supposedly adding 12dB gain which would account for the apparently lower noise floor and smaller spike.

Paul

Check the spectrum from the AD alone, perhaps that's the bad boy.

Samuel

Strangely enough I recorded the sound of rain from my window a couple of days ago. When I looked at the spectrum the next day, I saw a sharp spike at pretty much the same frequency. I suspected the DC converter circuit in the mics to be the culprit, but when I made a test with the same mics, same preamp and everything, the spike was gone. :? I'm beginning to think that the spike maybe was caused by something that emitted a 15+ K sine, perhaps a car alarm system or some device supposed to chase away insects???

My thought on this is thus,

The first graph shows the spike that is only 2 dB above what you have going on there. No big deal, imo. The second shows it about 4dB above...now I'm curious.

But seeing that its built on perfboard and with no case makes me to say: at least put it in a case. Even so, you may get conducted trash thru the cables connected to a A/D or it might be RF interference. What did you do exactly about RF filtering on the input and on the power?...yeah the batteries...bypass caps for the battery powered preamp...don't leave home without it.

More interesting to me would be a spectrum of a recording with the inputs shorted to gnd. That would tell you something for certain.

HTH!

I should ditch at least one of those graphs to avoid confusion. The two spectrums are made from recordings made on different devices - first one was made on a Apple G4 laptop using built in "sound card" line in. The second was made using the Sony RH1 HiMD recorder. The differences seem to reflect gain applied - the G4 having less gain that the RH1.

There is some RF protection - 2 x 470pf caps to 47pf to ground as per THAT1510 datasheet, and a 4.7nf cap between +/- signal just right at the THAT1510. I'm considering replacing the caps with a CM choke as has been suggested in another thread.

The perf board setup is just trial build to see how well the circuit worked, and to iron out any problems. I'll layout a pcb for a battery powered, 2 channel + phantom power version once I've ironed the main issues.

I'll try grounding the inputs and recording into the RH1 tonight and post the results.

cheers
Paul

There is some RF protection - 2 x 470 pF caps to 47 pF to ground as per THAT1510 datasheet, and a 4.7 nF cap between +/- signal just right at the THAT1510. I'm considering replacing the caps with a CM choke as has been suggested in another thread.

Click to expand...

A CM choke does not replace a cap, it's just an optional addition to caps.

Input RFI protection is not the problem here--that comes into play with 100 m mic cable in front of your pre. Get a decent case, there's not much point spending more time with this problem before you haven't gotten a basic shield around your circuit.

Samuel

If the interference is at exactly 15,734 Hz, you've just hit on someone's television set displaying NTSC. I didn't see what country you were from. PAL and SECAM use 15,625 Hz.

Television sets generate a lot of EMI. After all, there's a rather large electromagnet being excited with these frequencies in there. You could also be the innocent victim of EMI from a nearby television transmitter... in any case, right above 15 kHz is the TV horizontal scanning frequency.

Shielding might help. If you're getting magnetic interference you might need soft steel or mumetal - aluminum might not get it all.

Following Samuel's sage advice I tried putting the the board into a old folded steel project case, and used some short jumper leads to tie the xlr jack to the chassis. This seems to have cleaned up the interference spike with the 150 ohm resistor in place.There is a tiny spike of perhaps 0.5-1.0dB at -76dB with the +/- grounded to chassis.

My girlfriend just pointed out that there is a TV antenna at eye level about 10 meters away, on the roof of the apartment next door, so TV interference is probably a good call.

I'm in Australia, which uses PAL and - surprise, surprise - the spike was pretty much dead on 15626Hz.

thanks for being so patient will a newbie

cheers
Paul

Its not a problem to help anyone of any experience level when they can provide relevant information and carry out the suggestions given!

Good call Dale on the TV interference freqs!

I've attached a frequency analysis of the "in case" pre->rh1 line with 150ohm terminating resistor. Amazing what a bit of shielding achieves.
:wink:

Glad you found it!

For a bit of a laugh, NTSC = Never Twice the Same Colour....

[quote author="pj"]with the rest of the circuit taken directly from Jung and Garcia's Analog Devices AN-242 app note.[/quote]
Thanks for the URL, I hadn't seen that AN-242 before.

A bit puzzled though by the two anti-series caps for the gain-resistor.... The text states that these electrolytics are properly biased since the IC-pins are some 0.6V negative.... but the gain-resistor is floating :roll:
An added 1M resistor could arrange this, but as is it doesn't seem complete in this respect.

FWIW: w.r.t. the phantom-power & series-resistors (47 each IIRIC), interesting idea to 'remove' (say short) these and indeed, they won't harm when there's a hot condenser-mic in front. Dunno if that very switching (to say a phantom-less config without those series-resistors) is guaranteed to be plop-free.

Regards,

Peter

Hi Peter,

Luckily someone had a copy archived on their site - it's the only place I've seen it. The same circuit with dual OP176 opamps instead of the OP275 is discussed in Jung's "Microphone Preamplifiers for Audio: A brief guide for users" published in Analogue Dialogue 28-2 (1994) pp12-18. http://www.analog.com/library/analogDialogue/archives/scans/AD28.2-0894-FOA.pdf This article has a fantastic section on system RFI prevention.

The caps around RG are intended for use when you are using adjustable gain, so I suppose would be useful in blocking dc spikes when switching between fixed gain levels. They aren't required with a single fixed resistor, however I have a stepped attenuator almost ready to hook up.

I agree that the idea of switching all the phantom components out of the circuit, rather than just killing power is a great idea. I've been thinking about how best to implement this when I do the PCB.

On the subject of case designs, I was wondering if anyone might have pics that show internal case design of wendt X2/X3 or shure FP33 mixers??

cheers
Paul

I've used the lower values from the THAT1510 datasheet for the resistors. I have 10R in at the moment and will pick up some 4R7 when I do an order from a local surplus barn. I've had to use standard 1N4004's as I haven't found a decent local source for the 1N4004GP's unfortunately.

The blocking cap are pretty average quality too. I'll put some Panasonic's in when I get the boards sorted.

cheers
Paul

If you have a AC coupling capacitor in the feedback network you don't need to worry about diode leakage. That's a matter for DC coupled designs, but here you can use ordinary 1N4004 without any selection.

Samuel

Perhaps, never thought of that. Any data on how leakage diode current looks like?

Samuel

Surface leakage tends to be the problem at low frequencies and is highly package/process dependent. Bulk, before avalanche sets in, is usually pretty white and Gaussian.