I've been looking for a decent tone control preamp and have tried a couple Chinese boards that were awful. I was considering the ESP HiFi preamp but this was deemed a poor design as discussed on a different post.
This is another board kit I found which has decent specs. The balance control was added by me but not sure if that's the optimal location.
This is another board kit I found which has decent specs. The balance control was added by me but not sure if that's the optimal location.
Last edited by a moderator:
On the first op amp, the gain is wrongly set.
The feedback resistor should be the larger one, and the resistor to ground is the smaller one.
But their ratio is typically around ten rather than 100.
The balance control cannot work as shown.
The feedback resistor should be the larger one, and the resistor to ground is the smaller one.
But their ratio is typically around ten rather than 100.
The balance control cannot work as shown.
The balance circuit shorts the output of the opamps at the ends of the potentiometer range, and since the output impedance of an opamp is measured in milliohms (at low frequencies at least), the balance pot is doing nothing but overloading those opamps at the ends of the range. Perhaps the balance pot was originally connected to the wipers of the volume pots, which makes sense.
Modern low-noise design typically uses lower resistance values, like 10k pots, to help push the noise floor as low as possible.
Using a Baxandall volume stage at the input would cancel the inversion of the tone-control stage, but would require a difference balance circuit.
Modern low-noise design typically uses lower resistance values, like 10k pots, to help push the noise floor as low as possible.
Using a Baxandall volume stage at the input would cancel the inversion of the tone-control stage, but would require a difference balance circuit.
Thanks. I changed the balance to use the ESP better balance control and increased the gain to something closer to what I'll probably need.
Why does it matter if the signal is inverted?
Attachments
Some people can definitely hear inverted polarity in loud low frequency transients - think cannons going off in Tchaikovsky's 4th, or loud kick drums. Professional equipment is scrupulous to avoid inversion as in a mix you never want to accidentally combine a signal with an inverted version of itself (causing unwanted cancellation).
I think you took the wrong idea from that thread/post. The "poor design" opinion given of the ESP was about the linear pot with the log faking resistor (it's got it's plusses and minuses...), but not the ESP preamp itself.
This design you are looking at now may be argued as being an actual "poor design", (or at least poorer) for the myriad reasons the above posts in this thread point out.
Since I already have the ESP board I will go with that then.
It was a pretty withering critique though, and not just for the volume control.
It was a pretty withering critique though, and not just for the volume control.
If you re-read that thread carefully, you will realize that there's absolutely nothing wrong at all with the ESP pre. The critique was leveled at the idea of replacing log pots with loaded linear ones willy-nilly, which can indeed lead to problems, but there's nothing willy-nilly about this well thought out implementation by a competent designer who has considered those potential problems and addressed / avoided them so there are none left.
Consider eliminating the balance control which is not so useful as the bass and treble.
Add a bypass switch so you can totally remove any tone control effects.
The boost and cut of the bass control is minimal since Avb = 1 + RV2b/R19 ; 50k/10k; = 5
then 20 * log10(5) = 13.9db
I mention this only because a lot of WWW examples intend 20db (which might be excessive?)
Add a bypass switch so you can totally remove any tone control effects.
The boost and cut of the bass control is minimal since Avb = 1 + RV2b/R19 ; 50k/10k; = 5
then 20 * log10(5) = 13.9db
I mention this only because a lot of WWW examples intend 20db (which might be excessive?)
It is. I wouldn't consider it a major design flaw but, although many sources already have DC blocking caps at the output, there's no guarantee that will be the case and it's always a good idea to have them at the input of a preamp for protection. I'd add something like a 1u film cap and another 100k drain resistor to ground before it, which lowers the input impedance to 50k, still mild enough for pretty much any source.
Edit: I see in the picture a couple of electrolytic caps (the larger ones) marked C102/C202 which are nowhere to be found in the schematics, maybe those are the input caps and he forgot to include them in the diagrams?
Edit: I see in the picture a couple of electrolytic caps (the larger ones) marked C102/C202 which are nowhere to be found in the schematics, maybe those are the input caps and he forgot to include them in the diagrams?
Last edited:
Good catch, I forgot about those. They are on the revised schematic that is available when you buy the board(so I won't post it). They go between the inverting input and R103/203. So that would block any DC path to ground but without having to be in the signal path?
I vote for ESP's preamp though I would certainly replace the flaky LM4562s with NE5532.
But if you use his Balance Control, the pot has to be Wirewound else crosstalk would be excessive.
But if you use his Balance Control, the pot has to be Wirewound else crosstalk would be excessive.
Oh well, the 4562's (LME49720's actually) are already soldered in.
Why would wirewound have less crosstalk and is it better than a cermet?
Why would wirewound have less crosstalk and is it better than a cermet?
Non wirewound pots (including cermets) have significant resistance between the 'slider' and the track. A Balance pot in that position needs ZERO resistance between the earthed 'slider' and the track if you want no crosstalk
What you WILL have problems with is the DC reference coming though the level pot.
When it gets dirty (and it WILL) you will get some nasty noise. This is easily fixed
by adding a resistor from the "+" inputs of the input opamps to ground. You now
have a reference point regardless of the dirt in the pot. The value is not critical as
long as it doesn't alter the operation of the pot.
G²
10k:10k feedback network on the LM4562? That network contributes more voltage noise (9nV/√Hz) than the opamp (2.7nV/√Hz), no point using such low noise opamps with that arrangement really. And a lot of the other impedances are high enough that the opamp only needs to be < 6nV/√Hz to minimize overall noise - gives a wider choice of opamps.
But there is another issue which is that the LM4562 is bipolar and getting its bias current through the wipers of the various pots, which can lead to catastrophic noise bursts once the pots start to get even a tiny bit scratchy and start interrupting the bias current... Either modify the circuit with capacitors on all pot wipers (and alternate DC bias current paths), or choose a FET input opamp where the input bias currents are small enough to mitigate the issue. Otherwise you'll be having to change those pots much more frequently..
But there is another issue which is that the LM4562 is bipolar and getting its bias current through the wipers of the various pots, which can lead to catastrophic noise bursts once the pots start to get even a tiny bit scratchy and start interrupting the bias current... Either modify the circuit with capacitors on all pot wipers (and alternate DC bias current paths), or choose a FET input opamp where the input bias currents are small enough to mitigate the issue. Otherwise you'll be having to change those pots much more frequently..