Headphone amp + speaker sim problems!

[craigmillard]

hello all knowledgeable ones!

Wondering if anyone can spot a problem with the following headphone amp and speaker sim i have put together? Im getting one hell of an evil squeal coming through the headphones at the moment, From what i have read i can be an issue with the switching for the bipolar supply but i am using the max1044 which i thought should cut this out!?

Heres the schematic and board layout:





Any help would be much appreciated:)

Wanna get this on me pedal board for some late night practice!

Cheers
Craig

[Gurner]

That whole cct looks like a hammer to crack a nut.

If I'm reading it right, you're using a TL072 (IC22A/IC2B) to drive your headphones?

I doub't that'll work (headphones will typically have a lowish impedance & the TL072 isn't much good for driving low impedances)

But anyway, you've got to narrow the issue down...

Do you get a squeal still with the pot backed right off (feeding INTO IC2A/ IC2B) ....if yes, then that's eliminated 80% of your circuit - just a case then of trying to establish if it's the PSU (tip temporarily use batteries in place of the max1044) or the IC2A/IC2B circuit.

[ayayay!]

Did you include those jumpers?  Also might be time to Google "audio probe."  Good luck. 

[craigmillard]

hey guys, sorry didn't put much info on this:)

The opamp is actually 5532, i haven't updated the schematic!:(

I have separated the speaker sim from the headphone amp part and still get squeal through the headphones, using the aux ins to bypass the sim i can get audio but with the squeal, even if i turn down the volume for the speaker sim.

Would the position of the MAX1044 in the layout effect if the squeal gets picked up from the switching?

Cheers
Craig

[CynicalMan]

That looks a bit familiar...  

I have a few suggestions for the squeal.  If the ICs are socketed, I'd try removing the TL074. Also, does the squeal change if you change the bass and mid controls? What's the frequency of the squeal?

Why is there a current limiter on the supply voltage? Also, at which power rails is the squeal present (+9V, VA, VB, VC, VR)? If you have an oscilloscope that would be best for checking that, but you could just hook up an audio probe to them.

If you get this working, do you mind if I link to it on the Speaker Sim page? Having the link to the PCB layout would be nice. I should mention that I've modified the circuit a bit from the design that's on the page (I need to update that too). The modifications I made don't affect the sound at all, they just mean that you can use smaller parts. Specifically, I've replaced the 0.68u cap with a 68n cap and I've replaced 470 and 680 resistors that connect to it with 4k7 and 6k8 resistors. Like I said, this doesn't affect the frequency response of the circuit.

[craigmillard]

Hey Alex! Cheers for the Tips:)
Was considering making up a build document for this once I get it working, should be quite a nice little pedal board practice amp:)
I will PM ya the finished board (although my PCB designing skills are limited!) Great Speaker Sim Thou!:)

Yup taking the TL074 out does stop the squeal and the headphone amp part is fine!
Must be something funky going on in the Sim part then! I dont have my Audio probe here so cant trace it yet, hopefully get a look later on tonight!

The low R resistors and Caps are power filtering for my crappy switch mode power supply, not sure they are needed as the supply voltage is coming from the max1044 but thought id stick um in anyway!

Yer i did wonder about the 0.68uf cap, bit of a monster and i haven't got one at the mo so used a 0.47uf (still HUGE!).
I will update the schematic and board for your suggested changes thou, should make it easier to source parts for other people!

[CynicalMan]

I forgot to mention that the cap to Vr in the second op amp stage is now 680p. I've updated my site with the correct schematic.

I've been mulling this over, and I think the issue is one of the resonant filters. If there is enough parasitic feedback, filters like this will oscillate. I've had that problem with circuits like the ROG Condor and my G12H sim that have resonant filters, but it's always been pretty easy to fix up. I think the easiest way to test this out is to use an alligator clip cable to jumper each of the parts circled here one by one:


When you jumper the parts, they make the filter to which they're attached almost a buffer. Then, you can find out whether or not one of the filters is oscillating, and if so, which one is.

[craigmillard]

Hey there,

Been a while since i took a look at this!! Couldn't get it working so joined the pile of circuits in the didnt work pile!

Anyhow i have had a second stab at it and am still getting nowhere!

Did notice the c18 cap after the j2101 buffer was the wrong way round, so fixed that but no dif! Also tried shorting some of those suggested parts. Seems to cause it to work on occasion but cannot see why?

The Squeal is def coming from the speaker sim but cannot hone in on it!

Any more ideas or suggestions?

[CynicalMan]

Have you tried doing what I suggested in my last post? What were the results?

Also, what did you get with the audio probe?

[craigmillard]

Hey,

Almost a year to the day that i last looked at this! Cant believe its gone that quick!!

Anyways i have a working board and headphone amp finally!!:)

Thanks for the tips CynicalMan i redid the board with your 1.2 (minus the treble control) schematic and put the voltage divider network running from 9V instead of va. Great speaker sim sounds:icon_razz:

Only oddity now is a squeal that happens for about 3-5 seconds when first powered on! very odd! Also the squeal is only there if left powered off for 30 seconds or so and the plugged back in. Looks like a cap problem charging and discharging but not sure.

Any thoughts anyone?

Cheers

[craigmillard]

Anyone got any ideas on the squeal at power on? It disappears pretty quick but if you put the headphones on before powering its a bit loud
Can for the life of me see why?!

[Fergutor]

I have the same problems but for the Celestion G12H only version...well had because I'm dismantling this from the breadboard right now, which is a pitty because I wanted this as the base design for my own speaker sim in which I would have changed everything basically.  I tried what CynicalMan suggested, tried changing the capacitors to vref to ground, as I saw all other speaker sims with this kind of filters that way (while resistors to vref are to vref), see the Condor for example. I tried using an NPN based buffer. I tried changing the "order" of the opamps, I/vs used 2 tl072 IC1A > IC2A > IC1B > IC2B and then IC1A > IC2A > IC2B > IC1B just in case was something to do with the vref so it looked more like other speaker sims... and tried many other things that I don't remember now. The only thing that kind of worked is from 9v a 1n4007 and from that a 1n4148 to + of the first opamp (in) and another to ground (took that from the SE33 speaker sim), but it killed the bass. Also changing the electrolitic capacitor of the voltage divider made the squealing longer or shorter (less value shorter squeal).
Can anybody see what is wrong with this speaker sims design? https://sites.google.com/site/distorque/home/projects/celestion-g12h-speaker-sim  I have to try again after what's next in the breadboard.

[dschwartz]

I had the same issue with my last design (a super duper multicab sim), i almost lost my mind trying to find the reason why the darn thing would oscillate like a goddamn theremin when powered up..
After some google surfing and some reasoning, i finally found the cause...

Never, ever, design a sallen key resonant filter with a Q more than 3...it WILL oscillate..

The first lpf on the schem looks like it has a very high Q (judging by the 100pF cap) so you can expect a high pitched squeal..

Search in google for "sallen key calculator", find the cutoff frequency of your lpf, and redesign it with the same cutoff freq, but with  lower Q factor..2.5 or less..which isstill a good resonant filter, but well behaved..

[Fergutor]

I had the same issue with my last design (a super duper multicab sim), i almost lost my mind trying to find the reason why the darn thing would oscillate like a goddamn theremin when powered up..
After some google surfing and some reasoning, i finally found the cause...

Never, ever, design a sallen key resonant filter with a Q more than 3...it WILL oscillate..

The first lpf on the schem looks like it has a very high Q (judging by the 100pF cap) so you can expect a high pitched squeal..

Search in google for "sallen key calculator", find the cutoff frequency of your lpf, and redesign it with the same cutoff freq, but with  lower Q factor..2.5 or less..which isstill a good resonant filter, but well behaved..

Thanks! I will test when I have place in my breadboard. The first filter, of the verison I'm building which is almost the same as of the op, has a value of Q = 3.37099931232 , the second Q = 4.91592835656 (!!) the third, the high pass filter Q = 0.821583836258.  I will post my results later.
Hey Noisemaker why in your design (and many others) the filter caps go to ground and in this one to vref and all of them works? All those have single rail supply.

[PRR]

> filter caps go to ground and in this one to vref and all of them works?

Yes.

The filter caps need an AC(Audio) ground. Vref's bypass cap makes Vref "Audio ground".

The filter resistors which classically go "to ground" in this case must go to a point that is halfway up the DC supply (to center the opamps within the supply) yet is also a Audio ground. That is what Vref is for.