Variable trim in between 2 gain stage - ideal components value?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

elskardio

Well-known member
Joined
Sep 29, 2010
Messages
564
Location
Montreal - Canada
Hi Guys,

For the past week I’ve been testing a variable trim option for my mic preamp.

It’s a simple volume control, between the 2 gain stage, with an extra resistor to limit the range.

I’m presently using a 1K pot. I want to avoid any unnecessary load to the opamps. It’s working great and I didn’t measure any change in the noise level of the preamp.

Should I try with lower values?

Here’s the basic schematic:
Screen-Shot-2019-11-09-at-12-20-53-PM.png


Any suggestions?
Thanks
 
Lower would probably put unnecessary strain on the driving op amp and the noise improvement would be non-existent. I recommend using 10K.

Cheers

Ian
 
ruffrecords said:
Lower would probably put unnecessary strain on the driving op amp and the noise improvement would be non-existent. I recommend using 10K.

Cheers

Ian

Hi Ian,
Thanks for the response.
Isn’t 10k a bit high for the 1st opamp to drive?
 
elskardio said:
Hi Ian,
Thanks for the response.
Isn’t 10k a bit high for the 1st opamp to drive?

10k is a very common value for this purpose. Most modern opamps are happy driving down to 2K loads,  but not many will swing full output into 1K, and start distorting on a steep curve. Is that an issue for your design? See opamp output voltage vs distortion curves of  your device of choice.

That brings you to the basic question of what exact IC are you  going to use?  Going  higher with that pot is safe and will not be your first second or even third noise mechanism that contributes to the total noise figure.  The first stage will be the dominant source of noise, including all input circuitry. 

Even output voltage offset due to a high input current offset device that is aggravated by a high value pot is little concern if you are recording into an analog to digital converter.

To put your mind at ease look  at tube mic preamps that have 250k pots or higher between stages and they seem to be quit enough for studio recording when used with condenser microphones.

 
A decent op amp such as the 5532 and similar will have no problem driving the values you show (assuming the series feedback resistor is no lower than about 2K) but without a more complete schematic showing types used and gain distribution etc., it's a little hard to give an opinion on an optimised way to do what you want.
As long as the 1st amp can drive the combined load without ill effect,  I personally see no benefit to using a higher than necessary gain trim pot such as 10k, but that's just an opinion, please take it as such 😊
 
Winston O'Boogie said:
As long as the 1st amp can drive the combined load without ill effect,  I personally see no benefit to using a higher than necessary gain trim pot such as 10k, but that's just an opinion, please take it as such 😊
On the other hand there is no point in making it lower than necessary either. With op amp output stages being class B there will be a power supply current dependent on signal level and the load.  So a 10K load will draw 10 times less current and add ten times less crud to the power rails than a 1K load.

Cheers

Ian
 
ruffrecords said:
On the other hand there is no point in making it lower than necessary either. With op amp output stages being class B there will be a power supply current dependent on signal level and the load.  So a 10K load will draw 10 times less current and add ten times less crud to the power rails than a 1K load.

Again, without seeing a more complete schematic, it's hard to know which is the better way to go in terms of trade off. 
Of course, I agree with you re. the potential issues of a lower load.  But, I ASSumed that the power supply is up to task in terms of current, and that there will be proper bypassing of amps with no "crud" getting into the audio ground.  Why bother otherwise right? 😜
Optimizing for noise in the first, and the second stage, is also important in a mic amp, and a dB here and there adds up to, or diminishes from, a  first class job.    If the 1st and 2nd stages are something like an LM4562, then loads down to 600 ohms are perfectly fine for stage 1 and noise will be measurably less using the shown values of 1K and 499R  than with scaling them up by 10 and presenting a (worst case) 3K75 source plus stopper R to amp 2 if that amp Is a similar type and has also been optimized for noise.   
I see no blocking cap for stage 2 and if we are indeed looking at a bipolar input device in that position then we may need to examine current noise.

Also, although a 10k pot is unlikely to be an issue, a high source impedance will cause an increase in distortion at higher frequencies with certain op amps. 

Edit: Although it may actually be better to use something like an active Baxendall level control for the circa +/-10dB Elskardio is shooting for,  I think any of this I'm saying is up for grabs without seeing the whole picture.  Therefore, who knows?  😄

 
elskardio said:
Hi guys,
If it helps, This is for my twin servo preamp. Both opamps are discrete 990s. No capacitors in the signal path.
Thanks

Yes it helps.  Then with a 990 op amp you are better staying as low as possible in impedance, just as you showed us.  The 990 likes low source impedance and also laughs in the face of output loads that other amps throttle and choke their guts at.
Good work 👍
 
elskardio said:
...Isn’t 10k a bit high for the 1st opamp to drive?

A "heavy load" is a smaller number of Ohms.

Most audio electronics have a MINimum number of Ohms they will drive well.

TL072 is good for 2K or 3k to *infinity*.

'5532 is fine 500 Ohms to infinity.

'990 will drive 75 or 150 Ohms to infinity.

A higher value of volume/gain pot invites more hiss, yes. But the main source of hiss in a mike amp is the mike (and mike-input transistors). Where you have the gain control is MUCH higher audio level. And hiss voltage in higher value resistors goes as square-root of resistance. Or tolerable resistance goes as square of gain. If you take 100 Ohm mike and apply 30dB (30:1) of gain, then an interstage resistance must be 100r*30^2 or 100r*1000 or 100k Ohms to raise the hiss. 10k is a perfectly acceptable pot value here. (Why it is so very common.)
 
PRR said:
If you take 100 Ohm mike and apply 30dB (30:1) of gain, then an interstage resistance must be 100r*30^2 or 100r*1000 or 100k Ohms to raise the hiss. 10k is a perfectly acceptable pot value here. (Why it is so very common.)

The 990 in biased fairly high in its input pair, I've never been able to successfully and efficiently juggle servos so as to directly couple a 10K pot into one while also keeping the output at zero mV or close.      Maybe if the series feedback R was higher than Dean recommended.  Whereas, distributing gain between both amps, or keeping any inter pot low Z was WAY less headache.

Edit: I do take your point and it's a very valid and correct one.  But folks don't always record with technically correct gain structuring.  Amp one at zero gain and output amp at 30dB gain for less perceived tone is frequently done in rock n roll and is something a designer has no control over. 


 
After a break, a cigarette, and a wee think - I realized that I probably have no real clue as to what makes a mic amp a really  first class one.

At one time, all I seemed to do was bloody mic amps.  The last time I went to AES, there were 9 on display from 5 different manufacturers that I had a hand in.  The one I really hated has sold about 10,000  channels last I heard, while the one I liked the best was discontinued after a year.

I apologize for my, sometimes, dog-with-bone approach.
Peace 😊
 
Thanks PRR for the detailed explanation.

No need to apologize Winston. It's all great infos thanks

Winston O'Boogie said:
Then with a 990 op amp you are better staying as low as possible in impedance, just as you showed us.  The 990 likes low source impedance...
I found this in John Hardy's 990 documentation:

The ability to drive lower-impedance loads is important for two reasons. First, the 990 can easily drive multiple power amps, or pots, etc., with less concern for cumulative loading. Second, the resistors, capacitors and other parts that are connected around the 990 to determine the function of the circuit can be scaled down to much lower impedances than those of a monolithic design. This can result in lower noise. Some monolithic op-amps are theoretically capable of very low noise performance, but they cannot drive low impedances without in-creased distortion or decreased headroom, compromising performance

 
CJ said:
what about using NFB to control gain instead of the pot?  would that mean less noise?
In the sense that, with the shown scheme, amp no. 2 is always working at the maximum set gain no matter how low we set the inter level control, then yes. 
 
elskardio said:
Thanks PRR for the detailed explanation.

No need to apologize Winston. It's all great infos thanks
I found this in John Hardy's 990 documentation:

This can result in lower noise.

The important word here is 'can'. In the first stage of a mic pre, taking every measure to reduce noise is worthwhile. Reducing the values of NFB components is a pert of this and in that respect the 990 is superior to any monolithic op amp I can think of.

Once you reach the second stage - the one after the pot - unless the noise performance is really really bad it will make not difference to the final noise. The reason for this is that the noise from the source, amplified by the first very low noise amplifier, will be much larger than any noise introduced by the pot or the second op amp.

For example, lets say we have a mic pre with just 30dB of gain. Give it a 1dB noise figure which means with a typical 150 ohm source its output noise will be about -100dBu. The noise due to a 10K pot in the circuit originally posted is close to -119dBu so it will not affect the final output noise at all.

Cheers

Ian

 
elskardio said:
But is there any advantages of using a 10k pot instead of my 1k?
Not really here where you are using discrete opamps. Just contrary, I would say it is slightly better to use lower impedance pot if it isn't physically placed on the PCB between opamps,  i.e. if it is placed on a front panel, for example, and connected with wires.
 
ruffrecords said:
Yes, 10 times less signal related crud on the power supply lines.

Ian, I  don't see how that is, unless we assume the feedback impedance around the 1st amp has also been scaled up 10 times.  But if the 1st stage amp has been optimized for noise, it should be pretty low and will mostly swamp out that 10 X less load. 
Practically speaking, taking typical values used in 990 circuits, the difference will be closer to  2 X less load if the network is scaled up. 






 
Back
Top