tube mic pre build

[deveng]

Gents,

Here's something I've been working on for a few months. It's working but still needs some massaging. First gain stage has more gain than necessary and it still suffers from a bit of hum. Wiring still is not committed and may need some movement to clean audio up a bit. Tested with RMAA and it looks pretty good already. Frequency response is +/- 0.5 dB from 10Hz to 35kHz. Total gain is close to 60dB. The tube circuit is nothing spectacular or new. Basically a simple 2 stage 12ax7 feeding a 12bh7 follower. Jensen JT115k input transformer and a Jensen JT10k61 output transformer wired series at 4:1. Distortion is max 0.3% at full gain and just under 0.2% at lower gains. I'd like to drop these numbers if I can with this amount of gain in 2 stages and the topology used. IMD is higher than I'd like as well but I'll check on this after I lower the overall gain and clean up the hum issue. It's a MosFET/zener regulated B+ at 275v, regulated 12v filament, buffered VU (see "VU meter buffer" thread). The relays are setup for phase, 20dB pad and phantom.

regards,
Jeff

 

[deveng]

I pulled the first stage cathode bypass cap and added a power supply decoupling cap (with regulated B+ I thought I could get away with it). Total gain now is roughly 54dB and THD+noise dropped to 0.2% at 3/4 to full gain and down to 0.1% at lower gains. Hum is significantly lower but still slightly audible so I still have some work to do here. I expect when I reduce hum even further the noise spec will improve. IMD+noise dropped significantly more than than the THD. Its still has plenty of gain and can bring an SM57 up to recording levels without pushing the gain pot up much past 1/3.

Just did a quick test vocal and its noticably warm and smooth. Next test will be trying it on a high end condensor mic!

Regards,
Jeff

 

[rafafredd]

hum...? How are you grounding it? I usually end isolating the shields from the audio ground with a 10ohms resistor to get rid of hum.

 

[deveng]

Star ground to B+ low, then single wire to chassis. Also, XLR's single wire to chassis. Didn't use grid stoppers, or grounded gain pot (s.b. grounded to chassis by front panel but I'll check how well!). Swapped 12ax7's with no change. I may try shielded wires to input (attached to gnd at one end only).

Regards,
Jeff

 

[walter]

That looks good. It sounds like you are going in the right direction. Are you testing with the lid on?

 

[analag]

Disconnect the input transformer and listen to it. Signal wires and voltage wire layout scheme looks a little sloppy to me in that area.

 

[deveng]

Both good points, thanks. I had planned to ground and/or disconnect the input to debug the hum issue but had not thought about putting the lid on. All testing so far has been lid off. The only power wires near the input transformer are regulated +24 volts for the relays but they are wrapped up around them. After disconnecting the input, I might just un-bundle some of the the wiring and use CJ's "chop stick" method.

analag, I may not be able to lower my tube board enough to continue using a 12bh7 vertically. It's very close to the lid. Rather than laying down the 12bh7, I'm considering using a 12au7 for the follower and increase the current up from about 5ma(now) to 10ma. I doubt I'll ever see any 600 ohm loads and this still won't be near the maximum wattage. Any thoughts on this?

edited: I just found and read the thread "12BH7/12BH7a subs?" So, it confirms my suspicion that this tube would work for light to medium loads. But then again I like powering my JBL monitors (very efficient) with 200w per channel!

Regards,
Jeff

 

[pstamler]

Try detaching the ground lead from the volume control and instead running it to the ground point of the next stage's cathode resistor. That has sometimes been quieter for me than a strict star ground.

Also, instead of running the reference zener string (and its decouplng cap) in the regulator to the star ground, try running it to the load's grounding point, which then connects to star ground. Ditto the regulators for the filaments -- try connecting their grounds to the load ground rather than star.

Not textbook, but it has worked often for me.

Finally, I wouldn't go 12AU7 in this application; I'd go for a 6fQ7 or 6SN7 as the CF, even if you have to mount it horizontaly.

Peace,
Paul

 

[deveng]

Thanks for the advice Paul. Haven't had a chance to get at this low level hum thing yet but I'm soaking in all the experiences before I dig in.

After doing some critical listening through this mic pre (ignoring the low level hum for now) I find it very interesting. I had heard many say they highly recommend against using cathode bypass at least in large values. I can hear why now. To maximize my gain I had used a large capacitor on the first stage. The THD was 30% higher and IMD was "too high" as well. The IMD seemed to be the big culprit and was particularly nasty sounding. Just dropping the THD from 0.3% and 0.2% running wide open was worth losing 6dB of gain. I had hoped to get even lower but I read Eddy Cilleti's tube comparison chart and it appears for my configuration,tube selection and gain this may be the best I can get. I am now using 20dB of feedback and if I need more gain I may choose this avenue to get a few more dB. Not for now at least.

I have spent most of my time with IC opamp and discrete designs where distortion levels are more of a technical challenge than a large part of the sound, but this mic pre is an ear opener. I highly recommend those who've not tried a tube mic pre, to build one. It's not going to be one of those units I use on everything but will surely be a tool that is needed in the arsenal. I can't wait to drop the low level hum and get this thing on some serious recordings!

Jeff

 

[analag]

12AU7 will work, though I feel more comfortable running it at 7-8 ma. There are a few more tube types you could use too.

 

[deveng]

Dug in today and solved the hum issue! I found that my input transformer was not totally isolated from the chassis which was the biggest culprit. Second, I added shielded Mogami quad cable from the input to 1st stage as well as to and from the gain pot. Also, since I'm not using a cathode bypass cap now, I had room to drop the PTP board down and the 12bh7 fits under the hood (barely). It's very quiet. Now that the low level hum is gone THD and IMD distortion dropped again as well. I can open the gain pot full open and there is no hum. Quiet as a mouse! Only one thing left to do now. Since the meter circuit still pushes the meter to full on power up/down, I'm going to build a timer controlled relay to switch in/out the meter.

Got this beast done just in time. I'm recording vocals for a project on Tuesday and this will get a workout.

analag: you were correct. It was input related, but not due to power supply wires. Got me looking in the right place though.

Regards,
Jeff

 

[deveng]

Update:

I've now recorded numerous vocal tracks using the new unit (both male and female with a number of microphones) It sounds stunning good. I have a rack of API style mic pres and they sound really great but for vocals this mic pre is outstanding. I designed and built this unit just as an "alternative" to my discrete mic pres. Now this is my "go to" mic pre for vocals. Haven't tried it on anything else yet but I surely will as the recording progresses.

Jeff

 

[rlaury]

Looks Great!
What was your source for the toroid?
Does it provide filament and phantom or just the HV.

RonL

 

[deveng]

Ron,

Thanks. It's a custom transformer built to my specs. See my Black Market post here:

http://www.groupdiy.com/index.php?topic=22126&highlight=

Boris at Engineering Vista was great to work with and the prices are very reasonable.

The transformer is a triple secondary providing:
Sec. 1: 260V @ 200mA (B+)
Sec. 2: 44V @ 150mA (phantom and meter buffer)
Sec. 3: 12.6V @ 1.5A (filament and relay)

All voltages are regulated. The B+ uses a power mosFET and a diode string to set the voltage. Very low ripple. Filament, phantom and meter buffer supplies are typical 3 terminal regulator designs.


jeff

 

[BYacey]

The 12AX7 could probably benefit from a Mu metal can as well.