Hi everyone,
I built a LM1875 guitar amp and it works great except for the thump when the relay kicks in.
Ive used a 555 to delay the turn on and ive even got it out to 8 seconds as a test and it still happens. The 2 HPF / LPF switches also have small pops when they are changed.
I assume since the output cap has ample time to charge up i have either a very leaky capacitor or another source of DC getting to the output. Is there anything blatantly obvious that you guys can see here that could be the cause? I'm going to redesign my PCB so any notes would be great.
PS, That reversed cap at C9 is installed the other way, despite the schematic 🙂
I built a LM1875 guitar amp and it works great except for the thump when the relay kicks in.
Ive used a 555 to delay the turn on and ive even got it out to 8 seconds as a test and it still happens. The 2 HPF / LPF switches also have small pops when they are changed.
I assume since the output cap has ample time to charge up i have either a very leaky capacitor or another source of DC getting to the output. Is there anything blatantly obvious that you guys can see here that could be the cause? I'm going to redesign my PCB so any notes would be great.
PS, That reversed cap at C9 is installed the other way, despite the schematic 🙂
Add a 470ohm to 1K resistor from negative side of C14 to ground to let it be charged. 2 second delay will suffice.
Oh god lol. I assume that the cap isn't actually charging up as the negative leg isn't actually connected to the ground until its relay actuates? What a silly oversight =/
AAs for the small pops on the filter switches, is there anything to be done about them?
No, in the form the relay is wired, the negative side of C14 is floatant, so it establishes its normal VCC/2 value when relay puts in service the speaker. Another way is to reverse relay wiring putting the arm to C14 and in the NO contact the speaker and in the NC a 100R resistance for example, depending on if the relay is activated or deactivated after delay.
By the same reason, use a 100K to 1M resistor paralleled to C3 to attack the other smaller poping.
By the same reason, use a 100K to 1M resistor paralleled to C3 to attack the other smaller poping.
Thank you. Just a query. Why does c3 cause popping and it's never disconnected from ground, is it the momentary disconnect from the switched resistors? Also, one of C5 and C6 are left floating when switched. Same deal there too? Sorry for the questions
Every join between two metals creates a potential per se, more over when they are subject to increased temperature (Thermocouple) or chemical agents (Batteries and cells) called contact potential. This effect is unavoidable. If your amp is capable of amplify it, you will get a sound.
This contact potential was used in Tube circuits as source of bias and in certain circumstances can measure up to a volt.
Over it; as we are surrounded by RF from.several frequencies and field strenght, any unlinear element capable of rectify it will create small potentials elsewhere. See for example Energy Harvesting of the Linear Technology brand.
This contact potential was used in Tube circuits as source of bias and in certain circumstances can measure up to a volt.
Over it; as we are surrounded by RF from.several frequencies and field strenght, any unlinear element capable of rectify it will create small potentials elsewhere. See for example Energy Harvesting of the Linear Technology brand.
Last edited:
Its not the best of circuits.
The circled op amp shows none inverting going to 16v ! it should be pulled to half VCC using resistors if single pole power supply.
The output is floating so cap will have floating voltage discharged into speaker.
The circled op amp shows none inverting going to 16v ! it should be pulled to half VCC using resistors if single pole power supply.
The output is floating so cap will have floating voltage discharged into speaker.
Not sure if I've interpreted your reply correctly but as I said in my original post, I realised the cap polarity after I made the PCB with the polarity of the actual cap is correct ( soldered backwards to the orientation shown here). The 16v is half of the VCC to the amp which was 32.
I've learnt I don't need to regulate for the amp as it has a high psrr, so that will be changed.
The floating output cap was identified in the first reply.
The purpose is to learn and identify what errors were made, so if I've missed something let us know. Thanks 👍
The input filters shown are not very useful.
* Switching between R2 and R3 "does nothing" since the filter is driven by MUCH higher Guitar pickup impedance anyway.
Simulators/calculators will happily tell you the different cutoff frequency between 5k/10k and C3, Real World says that you have at least 50k in series with that, and that your guitar will sound dull, no sparkle, real world situation is you have huge 1500pF in parallel with input.
Normal practice is having 100pF there, no more than 220pF, tops.
* same with C5/6/7 switching
They are in series with 1M, so all you are doing is switching between 80Hz and 49Hz (weak) cutoff, way too low (for a Guitar) to have any audible effect.
* And in general, it being a high impedance weak signal source, it is good practice NOT to use any filtering at the input of a preamp, straight driven by a guitar pickup.
Best is to offer guitar a solid flat high input (470k-1M) to make it happy, at least use a buffer there for separation, and later add what you want.
Most you want there is 100-200 pf across input (and that after a 10k to 47k resistor) to tame RF interference, but nothing higher.
* Switching between R2 and R3 "does nothing" since the filter is driven by MUCH higher Guitar pickup impedance anyway.
Simulators/calculators will happily tell you the different cutoff frequency between 5k/10k and C3, Real World says that you have at least 50k in series with that, and that your guitar will sound dull, no sparkle, real world situation is you have huge 1500pF in parallel with input.
Normal practice is having 100pF there, no more than 220pF, tops.
* same with C5/6/7 switching
They are in series with 1M, so all you are doing is switching between 80Hz and 49Hz (weak) cutoff, way too low (for a Guitar) to have any audible effect.
* And in general, it being a high impedance weak signal source, it is good practice NOT to use any filtering at the input of a preamp, straight driven by a guitar pickup.
Best is to offer guitar a solid flat high input (470k-1M) to make it happy, at least use a buffer there for separation, and later add what you want.
Most you want there is 100-200 pf across input (and that after a 10k to 47k resistor) to tame RF interference, but nothing higher.
That's a lot of stages of AC coupling through RC filters with and without bias. Any and all capacitors have frequency responses determined by the impedance of the circuit before and after. I have no idea how to calculate all of it so I used LTSpice when I did mine. Small changes matter. Small caps tend to pass higher frequencies and block lower ones, by how much depends on the circuit.
Using a DC voltage rail as a direct VGND bias is surely going to couple power rail noise into the ground and have the potential for the ground bias to move due to current load.
I reinforced my VGND with a high current opamp and 400uF caps on both sides of it. So that when the opamp takes a big gulp of current there is ample charge in those caps to provide it and provide a low impedance path to ground. Maybe you have already enforced it upstream in the PSU, I would suggest further reinforcement locally. People so often forget about the return path and that the VGND carries real current, in both directions.
On electrolytic on the output. Consider they are polarised. Consider what you connect this device to. Speakers/phones, fine. If you connect it to another audio device which is floating at a higher DC offset you will end up with the cap reversed polarity. You are floating yours at local +12V, the next bit of gear could be floating at 24V from a 48V power rail. Your cap has 12V on the + side and 24V on the - side. Technically the same thing happens with all AC signals through electrlytic caps and it's a YMMV thing, however the amount of current you are putting through it and how far the voltage is reversing will determine how long that cap lasts. A technique is to use 2 electrolytic back to back. Again it depends on the circuit and the current. If you want bass output you are kinda stuck with electrylitics though. I suggest over specing them as far as your comfortable with.
Using a DC voltage rail as a direct VGND bias is surely going to couple power rail noise into the ground and have the potential for the ground bias to move due to current load.
I reinforced my VGND with a high current opamp and 400uF caps on both sides of it. So that when the opamp takes a big gulp of current there is ample charge in those caps to provide it and provide a low impedance path to ground. Maybe you have already enforced it upstream in the PSU, I would suggest further reinforcement locally. People so often forget about the return path and that the VGND carries real current, in both directions.
On electrolytic on the output. Consider they are polarised. Consider what you connect this device to. Speakers/phones, fine. If you connect it to another audio device which is floating at a higher DC offset you will end up with the cap reversed polarity. You are floating yours at local +12V, the next bit of gear could be floating at 24V from a 48V power rail. Your cap has 12V on the + side and 24V on the - side. Technically the same thing happens with all AC signals through electrlytic caps and it's a YMMV thing, however the amount of current you are putting through it and how far the voltage is reversing will determine how long that cap lasts. A technique is to use 2 electrolytic back to back. Again it depends on the circuit and the current. If you want bass output you are kinda stuck with electrylitics though. I suggest over specing them as far as your comfortable with.
Last edited:
As you seem to have an abundance of power supply rails, have you considered just using a proper split rail? It will make most of the GND/VGND issues go away entirely. You don't need to create a DC offset, because you are referencing actual ground with a + and - voltage rail either side across the opamps.
Sorry, just for context I'm using it as a power amplifier with the input being fed by a Line 6 helix with a line level signal rather than the guitar directly.
I'm on my phone and running off 5 hours sleep so I'll revisit this later and reply properly.
Thanks for the reply. There is a seperate PCB with power filtering and regulation on it. I had considered using a split supply, but I had a meanwell SMPS which was enough for the small output wattage I wanted.
Again I'll revisit your post when Ive woken properly 🤣
Ok, had a coffee and a look.
The schematic posted is a bit misleading sorry.
1) the source impedance is essentially a line level preamp, so the Z is low, but I value your opinion on the filters, so I'll look into that more. maybe I need to look into the interaction with impedance and equivalent resistance.
2) the HPF actually uses a On On On toggle switch, so the cutoffs are 50, 80 and 150 hz. Although I can get rid of the 50 as there's no real Audible change.
Would you still suggest having the filters after the buffer? Or seperated by another buffer stage.
Also, is C8 needed as C9 will block the DC. I think for some reason I was worried the volume pot would change the output z if the TL buffer, but I now think I'm wrong
Well, IF it will be driven by Preamp things change.
But unless otherwise stated, I must assume this "LM1875 Guitar Amp" will be driven straight by a guitar 🙂
Back to design choices: in principle Guitar amps are not "subtle" by any description, and EQ is always "noticeable", to put it mildly.
For example, famous VOX AC30, which defined 60´s British Rock sound, uses a 500k volume pot fed through a 500pF capacitor.
In modern values: 470pF and 470k, a BRUTAL cutoff of everything below 700Hz !!!!!! 😱
Is that described as a tinny weak anemic amplifier?
Not at all, but "bright and jangly", the Shadows/Beatles/Stones/Status Quo/Queen sound !!!!
Listen at those guitars!
Remember those are fat sounding Gibson ones, not bright squeaky Fender, and yet ....
I congratulate you on rolling up your sleeves and designing/building your own; I also suggest you have a strong hard look at as many as possible successful commercial amplifiers , and check resistor/capacitor combinations along each one ... you will be surprised.
A great learning tool.
But unless otherwise stated, I must assume this "LM1875 Guitar Amp" will be driven straight by a guitar 🙂
Back to design choices: in principle Guitar amps are not "subtle" by any description, and EQ is always "noticeable", to put it mildly.
For example, famous VOX AC30, which defined 60´s British Rock sound, uses a 500k volume pot fed through a 500pF capacitor.
In modern values: 470pF and 470k, a BRUTAL cutoff of everything below 700Hz !!!!!! 😱
Is that described as a tinny weak anemic amplifier?
Not at all, but "bright and jangly", the Shadows/Beatles/Stones/Status Quo/Queen sound !!!!
Listen at those guitars!
Remember those are fat sounding Gibson ones, not bright squeaky Fender, and yet ....
I congratulate you on rolling up your sleeves and designing/building your own; I also suggest you have a strong hard look at as many as possible successful commercial amplifiers , and check resistor/capacitor combinations along each one ... you will be surprised.
A great learning tool.
Last edited:
Thank you very much! I just had a look at the AC30 and that 700hz cutoff is CRAZY!!
I think ill be delving into a lot more theory to get a bit more of a in-depth understanding of it all. Just for flexibility would it be best to have 2 buffer stages with the filters in between incase I went straight in with a guitar or another hiZ input?
Also, if I were to remove that output cap from the buffer, would the Volume pot have any influence on the output Z of the opamp stage going into the power amp?
I think ill be delving into a lot more theory to get a bit more of a in-depth understanding of it all. Just for flexibility would it be best to have 2 buffer stages with the filters in between incase I went straight in with a guitar or another hiZ input?
Also, if I were to remove that output cap from the buffer, would the Volume pot have any influence on the output Z of the opamp stage going into the power amp?
