Mic Preamp with unshielded input transformers.

[Heikki]

Here's a mic preamp with low noise and hum even without shielded input transformers. DC-DC converters are used in the power supply to make B+ for the tubes and +/- 15V for the op amps. Input transformers are Neutrik NTE10/3.

Schematic
https://www.ghr.fi/proaudio/micpre/micpre4.pdfTransformers and and 1:3 / 1:10 switch not shown on the schematic.

Clip indicator schematic
https://www.ghr.fi/proaudio/micpre/clipindicator.pdf
Noise at 61 dB gain, 200 ohm resistor across input XLR pins 2 & 3
https://www.ghr.fi/proaudio/micpre/micprenoiseat61dbgain200ohm.jpg
I think that's pretty good with unshielded transformers since hum doesn't really ruin the noise figure. Most commercial tube preamp specs don't say anything about noise, so I can only assume this is better than most.

Here's a song we recorded yesterday using the preamp.
http://ghr.fi/audio/instrumenthol_r1_session.flac

 

[ruffrecords]

That is a very good result for an ECC83 and an unshielded transformer. In my experience, hum and its harmonics does materially affect the noise figure so I think you could probably do 3dB better with a screened transformer.

Cheers

Ian

 

[Bo Deadly]

Cool.

You should try to figure out where the hum is coming from. Not that it's a problem for a tube rig but it would be good to know the source.

It seems to me with it being entirely SMPS and the input is just a 200 ohm resistor and the output is fully balanced, it should be completely hum-free and year clearly that is mains hum.

I can only conclude that either the usual ambient electromagnetic radiation is getting pickup by the transformer or the high impedance / high gain bits or there is a problem with earth grounding or there is an issue with the recording setup.

 

[volker]

Nice work! Did you compare your circuit to conventional cathode bias? As far as I know grid leak bias usually results in slightly higher noise, would be interesting to know by how much.

 

[Heikki]

I can only conclude that either the usual ambient electromagnetic radiation is getting pickup by the transformer

You are correct. Shorting the transformer secondary gets rid of all the mains hum and it's harmonics.

That is a very good result for an ECC83 and an unshielded transformer. In my experience, hum and its harmonics does materially affect the noise figure so I think you could probably do 3dB better with a screened transformer.

For me 3 dB improvement isn't worth the 10 times higher price of a screened transformer.

Nice work! Did you compare your circuit to conventional cathode bias? As far as I know grid leak bias usually results in slightly higher noise, would be interesting to know by how much.

I actually have a almost identical version of this mic preamp where the only difference is that the tube stage is cathode biased. There is a marginal difference in noise but nothing significant. I'll make some measurements sometime this week and post them here. There is a difference in distortion compared between grid leak bias and cathode bias. THD is pretty much the same but grid leak bias has more 3rd harmonic. In my opinion the grid leak bias sounds better especially when recording drums but it probably doesn't have much to do with the distortion harmonic components.

 

[mhelin]

The Recom DC/DC converter for B+ is quite an expensive part (~50 EUR/USD), couldn't it be replaced with a some kind of CMOS inverter multiplier (if you parallel inverters you will get more current out of multiplier circuit)? If you are going to build several channels though then the price will be no issue... B+ 150V?

 

[ruffrecords]

For me 3 dB improvement isn't worth the 10 times higher price of a screened transformer.

You can buy a screening can for less than the price of the transformer. I am sure you could fit the Neutrik into this one:

A262CAN | Transformer Screening Can | RS

In fact the combination is so attractive I might just try it as an option for the poor man's tube mixer.

Cheers

Ian

 

[Heikki]

The Recom DC/DC converter for B+ is quite an expensive part (~50 EUR/USD), couldn't it be replaced with a some kind of CMOS inverter multiplier

Since B+ is a little bit over 200V and the plate current for each triode is about 0.5 mA, something like fig 11 here would probably work.

You can buy a screening can for less than the price of the transformer. I am sure you could fit the Neutrik into this one:

A262CAN | Transformer Screening Can | RS

Thanks for the link. If I ever build another one I might try out those screening cans.

 

[Bo Deadly]

Put the transformers in a steel box. Fine an electrical junction box. Even if it has holes for screws and such, unlike electric field radiation which can get through any gap, electromagnetic does not get through small holes as easily. Especially mains frequencies. Find two boxes where one fits within another for two layers.

 

[abbey road d enfer]

For me 3 dB improvement isn't worth the 10 times higher price of a screened transformer.

A less costly option is to use two of these xfmrs in a humbucking configuration.

 

[Gavin Pursinger]

Interesting concept....humbuccking input vformers. Never thought of that. How would you go about doing this? Stacked right angles?

Thcx

 

[Bo Deadly]

Interesting concept....humbuccking input vformers. Never thought of that. How would you go about doing this? Stacked right angles?

Presumably they would just be wired using opposite wires. So if you have two transformers A and B each with primary wires 1 and 2 and signal source wires P and N, then you wire P to A1, A2 to B2, B2 to N. So at any moment the current is running one way through one transformer and the opposite way through the other. If the transformers are placed as close together as possible, any electromagnetic interference would induce currents in each transformer that are equal and opposite and therefore cancel each other out. Not sure how much the spacial orientation matters but I would guess orienting them in the same way would work best.

 

[Gavin Pursinger]

I have a pair of identical ,(?) xformers and will definitely experiment with your advice...like a humbucking guitar pickup. Awesome ...thx

 

[Heikki]

Nice work! Did you compare your circuit to conventional cathode bias? As far as I know grid leak bias usually results in slightly higher noise, would be interesting to know by how much.

Grid leak bias noise at 61 dB gain
https://ghr.fi/proaudio/micpre/gridleakgain61db.png
Cathode bias noise at 60 dB gain
https://ghr.fi/proaudio/micpre/cathodebiasgain60db.png

 

[VertigoRecording]

You can buy a screening can for less than the price of the transformer. I am sure you could fit the Neutrik into this one:

A262CAN | Transformer Screening Can | RS

Good one! Only $7.705 US Dollar.

 

[fazer]

Nice sound. Loads of fun without to much cost. Pcb s?

 

[jokeramik]

Which transformer do you use for input and output?
Do you measure the frequency response of the practical circuit with and without 600 Ohm load on output?
Please publish the pictures for the transmission of square-wave signals.
Best regards!
jokeramik

 

[fazer]

Input transformers are Neutrik NTE10/3.

Which transformer do you use for input and output?

In his first post. No output transformer. Open the schematic it uses DVR134 chip for balanced out.

 

[jokeramik]

In his first post. No output transformer. Open the schematic it uses DVR134 chip for balanced out.

No problem thanks and sorry for the output transfomer...)
Best regards!

 

[dmp]

Cool idea - missed this when first posted.
I wonder if using a differential receiver chip would be preferable to the tiny unshielded input Tx that is picking up hum?
DRV134 should have 1uF near the power pins based on the datasheet, vs 0.1uF
Also for a barebones preamp I don't think the 16 overvoltage protection diodes on the output are necessary? Or the clip indicator (but that is just my preference)
Just a few thoughts - great job.