Idea of an simple mic pre.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

hans a

Well-known member
Joined
Feb 17, 2014
Messages
60
Skärmavbild 2023-06-10 kl. 10.46.55.png

This is what i tinkered with. I tried making a test pcb, but it didn´t work. I wonder why that is.
If anyone has a thought please share.

Edit: the idea was to use this instead of the 4559 mic pre in tascam mod 3. It has a tamura input tx. I have substituted it with a resistor.
Br Hans
 
Last edited:
View attachment 109956

This is what i tinkered with. I tried making a test pcb, but it didn´t work. I wonder why that is.
If anyone has a thought please share.
Br Hans
That is not really a mic pre amp, that is just a common emitter amplifier with a Sziklai transistor arrangement, if you really want a mic pre you should start by having a differential input.

I don't know what that 1.7k resistor is supposed to be, is that supposed to represent the mic's output impedance, if so, it is too high, if not, why is it even there? It will only add noise, and you are attenuating the input signal just to amplify it back again, this will wreck the SNR.

Why did you use the bootstrapped 18k resistor? Makes no sense to me. Is the 6.8pF cap for stability? If so, what made you use it and how did you determine that value? This type of amplifier, that is, with a bypassed emitter, distorts madly with inputs slightly larger than 2-5 mV, since there is no emitter degeneration; although the Sziklai arrangement might help, still, it is extremely non-linear. The input impedance is around 4.2 K, which makes it unsuitable for most mics. The output impedance is around 6.8K, which makes it excessively high for line level interconnections (change that 100K resistor to a more common 10K, as used in line-level interconnections, notice the results). The gain is around 50dB, but this value will vary as a function of the input signal.

All that being said, it is a good effort, keep experimenting.
 
Thank you!
Gyraf: I have a tascam Model 3 and have been thinking (for fun) about bypassing the opamp of the pre and using very simple single-ended circuit. Also for me to understand circuits better. Actually it seems to work in simulation. I don´t know how the second transistor gets bias. I used the generic transistors in falstad. (Maybe theres a potential difference between gate and collector because of the voltage drop between the base and emitter pulling down the gate of the npn-transistor? im guessing here.)

user 37518: Thanks the points.
I´m not understanding how to calculate the input impedance correctly and will have to read up on that.
The 1.7k resistor is in place of a tamura input transformer in the Tascam Model 3. I found some numbers on ebay tho they may be incorrect. I went by 10k, 1k and 50hz and tried to tune the capacitor. My first attempt i think had better linearity but i could´nt find the save. The thinking was that the cap would give negative feed-back and the resistor path would give positive feed-back but i see now that that is´nt the case and that it acts more like a current drain. Thanks!
 
Last edited:
I had already started fiddling but i will keep that in mind.. tried and upped the voltage of the input to what i believe are the voltages of a SM57 trough a step up transformer.
 

Attachments

  • Skärmavbild 2023-06-10 kl. 15.11.11.png
    Skärmavbild 2023-06-10 kl. 15.11.11.png
    61.4 KB · Views: 8
Last edited:
I had already started fiddling but i will keep that in mind. I tried again and upped the voltage of the input to what i believe are the voltages of a SM57 trough a step up transformer.
That is better, I still think the 10k resistor is not needed. It is bootstrapped, which means that practically no current flows through it, bootstrapping is useful when you want to raise the input impedance, but not in that manner. The 22uF cap is too small, raise it to 100uF at least. The 6.8pF cap might not be needed, an engineering premise is not to add something which is not required. The Falstad simulator is very lame to model things like parasitic oscillations, build the thing in a protoboard and, if you notice oscillations, then you can think about adding that cap, otherwise, don't. Something important to keep in mind is that the 6.8pF cap experiences the Miller effect, with the gain you have, it presents an effective capacitance value of 172 pF, which is still ok, but, again, might be completely unnecessary. In your first circuit, since it had more gain, the cap had an effective value of around 3.3 nF due to Miller effect, which forms a pole with the 1.7k resistor at roughly 28 kHz, which is unpleasantly low, producing 1.8dB of attenuation at 20 kHz, 1.1dB at 15kHz, and 0.5dB at 10kHz; not good.

The 220 ohm resistor in parallel should give you better linearity, since it represents a degeneration factor of around 15, but it lowered your gain to 32 dB (not counting the 10K load), with the 10 K load the gain is 28dB, which is far from the maximum standard 60 dB gain of a typical preamp.
 
That is better, I still think the 10k resistor is not needed. It is bootstrapped, which means that practically no current flows through it, bootstrapping is useful when you want to raise the input impedance, but not in that manner. The 22uF cap is too small, raise it to 100uF at least. The 6.8pF cap might not be needed, an engineering premise is not to add something which is not required. The Falstad simulator is very lame to model things like parasitic oscillations, build the thing in a protoboard and, if you notice oscillations, then you can think about adding that cap, otherwise, don't. Something important to keep in mind is that the 6.8pF cap experiences the Miller effect, with the gain you have, it presents an effective capacitance value of 172 pF, which is still ok, but, again, might be completely unnecessary. In your first circuit, since it had more gain, the cap had an effective value of around 3.3 nF due to Miller effect, which forms a pole with the 1.7k resistor at roughly 28 kHz, which is unpleasantly low, producing 1.8dB of attenuation at 20 kHz, 1.1dB at 15kHz, and 0.5dB at 10kHz; not good.

The 220 ohm resistor in parallel should give you better linearity, since it represents a degeneration factor of around 15, but it lowered your gain to 32 dB (not counting the 10K load), with the 10 K load the gain is 28dB, which is far from the maximum standard 60 dB gain of a typical preamp.
Thank you for taking time and explaining. Learning a lot, and stuff that i need to learn more about. I will tinker some more.
 
Newmarket is completely right, the Falstad simulator is educational; not really meant to be used as a tool to check designs before you make a PCB.

Well, I wasn't meaning to be critical. But I've only heard of the Falstaf sim on this site. Maybe as I'm not in education. Professionally anyone I know uses some permutation of Simetrix/Simplis (paid for) ; LT Spice; TINA (mainly free TI version); TI Web Bench etc. ; ECAD integrated SIM eg Altium / MultiSim.
Falstaf always seems to be problematic in transferring to physical circuits ? Happy to be contradicted. No vested interest.
 
Well, I wasn't meaning to be critical. But I've only heard of the Falstaf sim on this site. Maybe as I'm not in education. Professionally anyone I know uses some permutation of Simetrix/Simplis (paid for) ; LT Spice; TINA (mainly free TI version); TI Web Bench etc. ; ECAD integrated SIM eg Altium / MultiSim.
Falstaf always seems to be problematic in transferring to physical circuits ? Happy to be contradicted. No vested interest.
The best Sims I've used, and still use (thanks to the University) are Keysight ADS and Keysight Genesys, particularly the former is in a whole different league, both in terms of price, power and accuracy. However, you know that any Sim is as only as good as its models. I believe I heard that you can load SPICE models into Falstad, but I am not entirely sure about this. I've know the Falstad sim since like 2008, and it has somehow progressed from being simply a cute Java applet, to being a more complex Java applet; but it is still a Java applet. Whenever someone shows me his/her circuits in Falstad, I know they are beginners.

A sim that I used to use back in my college days for digital stuff is Proteus, I no longer see many people still using it though.
 
The best Sims I've used, and still use (thanks to the University) are Keysight ADS and Keysight Genesys, particularly the former is in a whole different league, both in terms of price, power and accuracy. However, you know that any Sim is as only as good as its models. I believe I heard that you can load SPICE models into Falstad, but I am not entirely sure about this. I've know the Falstad sim since like 2008, and it has somehow progressed from being simply a cute Java applet, to being a more complex Java applet; but it is still a Java applet. Whenever someone shows me his/her circuits in Falstad, I know they are beginners.

A sim that I used to use back in my college days for digital stuff is Proteus, I no longer see many people still using it though.

Ah. My job before my current one had Proteus as ECAD. But not sure it had the SIM. It hadn't been maintained since 2014 or near. I switched new stuff to Pulsonix but never really got to use it for any real work.
But yes. Proteus seems out of the game now tbh.
 
To the OP (and everybody elese) I would recommend Micro CAP for circuit simulation/analysis. It's been available 100% free for some years now (unsupported thought) but it's a professional tool and not so hard to start running for the basic stuff.
I've heard a lot fo good stuff about Micro Cap, but with all the free sims available by big companies like TINA or LTSpice, I think it hasn't got much attention. However, there is a free simulator that is at a whole different level called Qucs-S. You should also download Xyce with it, since they work together. Qucs-S is a SPICE simulator but it goes beyond SPICE, it can do S parameters and Harmonic Balance analysis. You can couple it with Octave for co-simulation. I just started using it and I was blown away. It is like the free version of Genesys or something similar, minus the EM simulation.
 
Last edited:
Even simpler is a 1510/217/163 with a simple op amp output; how often do you reallly need a mic pre to have a balanced output? Unbalanced line levels with modern equipment and modern cables are quite noise-resistant.

I've routinely run them 75' and more with no noise at all.
 
Even simpler is a 1510/217/163 with a simple op amp output; how often do you reallly need a mic pre to have a balanced output? Unbalanced line levels with modern equipment and modern cables are quite noise-resistant.

I've routinely run them 75' and more with no noise at all.

tbh no good reason to not have an impedance balanced or ground sensing output. Still just one opamp stage. No point introducing an unbalanced interface when you're in control imo
 
Last edited:
tbh no good reason to not have an impedance balanced or ground sensing output. Still just one opamp stage. No point introducing an unbalanced interface when your in control imo
I must say that quite a lot of times I just prefer the "quasi-balanced" option of adding one resistor in series with the output for stability and an identical one from the cold terminal to ground.
 
I must say that quite a lot of times I just prefer the "quasi-balanced" option of adding one resistor in series with the output for stability and an identical one from the cold terminal to ground.
I've always been fascinated by this method used by Soundcraft; anyone seen it used anwhere else?
 

Attachments

  • Q-bal.png
    Q-bal.png
    25.4 KB · Views: 12
I must say that quite a lot of times I just prefer the "quasi-balanced" option of adding one resistor in series with the output for stability and an identical one from the cold terminal to ground.
Not to quibble about terminology but that topology is called "impedance balanced". Some remarkable bang for the buck when a couple cents worth of resistors allow the marketers to call the output "balanced" ;) . Not only that it works.
I've always been fascinated by this method used by Soundcraft; anyone seen it used anwhere else?
Impedance balanced with a simple differential output where the signal is referenced to the external ground. This (differential) is widely done for unbalanced insert points. The impedance balanced outputs are widely used on value products (because it's cheap). 🤔

JR
 
Not to quibble about terminology but that topology is called "impedance balanced". Some remarkable bang for the buck when a couple cents worth of resistors allow the marketers to call the output "balanced" ;) . Not only that it works.
Yes, however, if being completely anal, impedance balanced is whatever topology that has both impedances balanced with respect to ground.
 
Yes, however, if being completely anal. impedance balanced is whatever topology that has both impedances balanced with respect to ground.
In my experience in the mixer business (decades ago) "impedance balanced" was the terminology used to describe that very specific topology where you add one equal value resistor in series with the - output lead connected to ground to convert an unbalanced output to (cough) balanced. 🤔

There have been voluminous discussions about the different flavors of "balanced" right here. I recall years ago digging out my copy of the IEEE S-100 standards dictionary. That S-100 search was not very rewarding (I expect a search here could reveal some of our past discourse on this topic).

JR
 
Ime it's usually referred to as a "Ground Sensing" or "Ground Cancelling" output. In computational terms you can look at it as adding the destination 0V reference to the signal opposed to the input subtracting one input of a differential signal pair from the other. This makes it effective even when feeding an unbalanced input.
And if it feeds a standard balanced input then it acts as an impedance balanced output.
Both inverting and non-inverting configurations are covered by Self in "Small Signal...". Used by Soundcraft and, to my knowledge, DDA (and Dave Dearden of DDA was ex-Soundcraft) IIRC for some outputs on DMR24 and Profile desks.
 
If using an op amp or SSM2017/2019/'other' single chip amplifiers always remember to use a 'load isolating resistor 47 -150 Ohms or whatever value you want because hanging long cables on outputs can make them oscillate.
 
Edit: the idea was to use this instead of the 4559 mic pre in tascam mod 3. It has a tamura input tx. I have substituted it with a resistor.

The easiest option would be to upgrade the 4559 Op-Amp for something more modern and keep the rest more or less original.

Tamura Transformers have a good reputation. Change 4559 for OPA1662 or OPA1652, for example.
Thor
 
Back
Top