High Gain Amp Hiss Problem !

[PG000]

I am modding a JCM 800 2204(Clone) with the #39 Mod from Darkbluemurder(I attached the schematic). A few hours ago I solved one problem (mycrophonic squealing) but now I´m having another problem. The amp makes a very loud hiss noise and it´s get louder with the Volume turned up. Does somebody have an idea to solve this problem ?

PG000

[tubenit]

Do a search on the "enhance cap" which I use on all my LTPI amps. Do a search on snubbing caps.  Do a search on smoothing caps like they use on Dumble amps.

Another thing to consider is trying a 5751 in V1 position.  I typically can still get as much gain as I will ever want using a 5751.  And different 12AX7's will be noisier or quieter.  I think the Sovtek 12AX7's that Doug carries have been pretty quiet in my experience and good sounding for Marshallish type tones.

With respect, Tubenit

[PG000]

Thanks for your answer, I´ve tried enhance cap´s but they didn´t work out for me. I tried other 12AX7 but I didn´t recognize any difference. But I spoted the tube/s wich are "making" the hiss. It´s V1 and V2. Does someone have an idea how to treduce hiss at the first preamp stages(I am using shielded wire at the first 2 preamp tubes) ?

[VMS]

Does the hiss change if you chopstick the wires?

Did you solder grid stoppers to tube sockets?

[PG000]

No the hiss doesnt change if I chopstick the wires. I have grid stoppers before almost every stage like in the schematic.

[2deaf]

Four stage overdriven preamps are going to hiss.  This one compounds the problem by having three treble-boosting stages.  To reduce the hiss,  you have to attenuate the higher frequencies.  It becomes a compromise between how much treble you are willing to give up and how much hiss you are willing to tolerate.  Also, the hiss is a lot less noticeable after you blast your ears for awhile.

I recommend putting a 15pf to 22pf capacitor between the plate and grid of V1a right there on the socket.  For more treble cut, also put a 470pf to .001uf capacitor across R5 (100K) at V2a.  For even more treble cut, also put a 470pf to .001uf from plate to cathode on V2b.  You could also raise R11 (10K) to a six-digit value.     
 

[labb]

Using 1 watt or higher plate resistors sometimes helps a bit.

[FranciscoPerez]

I'd lower the first grid-stopper to 10K, and choose a good metal film 1W resistor for that position.

[trobbins]

Just focusing on input stage hiss generation, then apart from using something like PRO2 resistors for R14, R33, R3 and lowering R14 as suggested, then your next hiss issue is the 1 Meg gain pot, and the need for R4=470k to provide some gain dumping.  You could try using say a 250k gain pot, and 100k R4. 


Perhaps take some form of measurement when making changes, and adding tone shaping, to give a reference.  You can also use a spectrum analyser on an Android phone, or laptop to get a feel for the spectral shape of the hiss (it will depend on your speakers).

[VMS]

What did you do with the unused V1b half?

Also three gain stages sharing B5 node might be an issue.

[tubenit]

I am surprised that an "enhance cap" didn't work out for you?  It's always been a "must have" for me in killing unwanted harsh treble.  And the snubbing cap you're using has never worked for me.  I've tried it a handful of times and it always sounded like a blanket over my speaker and that it was a real tone killer.  I've probably never kept a snubbing cap in any amp more than 15 minutes.

However, I acknowledge amps are different and people do hear things differently. If you're getting the results you want then I'd stay with what you are doing.

Did you try a 5751 in V1?  You may find that you still have every bit of gain that you want still at hand.  OR the Sovtek 12AX7 tube?

Take a look at what these 4 gain stage amps do to lower hiss.

With respect, Tubenit

[tubenit]

It's worth noting that most of the 4 gain stage amps have fairly small coupling caps in at least one or more gain stages.   For example coupling caps as small as .0015 to .0047 in one or more gain stages.

In contrast, your amp has coupling caps that are .022 in C2, C4, & C6.  It may worth changing one of those to .0047 to see if that helps?
You could under one end of C2 and safely clip a .0047 into that position with insulated alligator clipped wiring.

With respect, Tubenit

[2deaf]

You could try using say a 250k gain pot, and 100k R4. 

Why?  That will squat the S-Shaped frequency response way down and move it to higher frequencies while lowering the overall gain a little.  I would rather just turn the knob on R15 down a little and retain the tone-shaping.

[PG000]

Thank you for all of your help. tubenit was suprised that the "enhance cap" didn´t work, so I tried it again and now it worked. I don´t know why it didn´t work the first time but now its great. I tried many different values and now I found the best. The hiss is now almost gone but now the amp is huming. Does anyone know a trick/fix to "prevent" hum ?

[shooter]

but now the amp is huming

Possibly you moved wires, chop-stick especially near V1, and any place you've recently been

[PG000]

I think i didn't move wires because I allready tried chopsticking the wires but nothing hapened. I think the hum was there the whole time but I didn't recognize it because the hiss was louder than the hum and now the hiss got less and I recognize the hum more 

[VMS]

Does the volume control affect the amount of hum?

[PG000]

Yes, the volume does affect the hum. The volume of the amp effects the hum a little bit but the Volume of the guitar effects it extremly.

[VMS]

It can be number of things, but here is a good debugging page that might help you:

http://www.geofex.com/ampdbug/ampdebug.htm

[trobbins]

You could try using say a 250k gain pot, and 100k R4. 

Why?  That will squat the S-Shaped frequency response way down and move it to higher frequencies while lowering the overall gain a little.  I would rather just turn the knob on R15 down a little and retain the tone-shaping.

The comment was about reducing resistor thermal noise, and perhaps contact noise from the pot as it is a grid leak, and its a pot (maybe the OP is using conductive plastic pot rather than vintage?).  Tone shaping could remain effectively the same by modifying the caps?

[2deaf]

The comment was about reducing resistor thermal noise . . .

Isn't the noise produced by an RC circuit completely independent of R?

Tone shaping could remain effectively the same by modifying the caps?

Raising C3 (500pf) to .0022uf when in parallel with 100K would probably yield a tone shape similar to 470K/500pf.  The noise voltage of an RC circuit may be inversely proportional to C, in which case raising C3 to .0022uf could reduce noise.

The noise coming out of the first stage is way larger than any noise R4, C3, and R15 could produce.  If you use the square root of the sum of the squares of each noise voltage to determine the total noise voltage, then the square of the first stage output noise voltage is going to overwhelmingly dominate the square of the input noise voltage of stage two.  The resulting noise at the grid of stage two is going to be essentially the noise coming out of stage one no matter what the values of R4, C3, and R15 are. 

[trobbins]

I agree that going past the first stage to alleviate circuit noise is likely of no benefit - but that said, an old 1 Megohm pot may be noticeable, so for trouble shooting it may be worth the effort to check.

[Ken Moon]

Regarding caps in parallel with plate resistors, whatever stage they're used on, isn't it better to connect the cap between the tube's plate and ground, vs the plate and HT?

Each cap shunts high frequencies to ground by implementing an RC low-pass filter (R being the tube's output impedance, and C being the bypass cap).

Either way you connect it, the frequency response is the same, since HT is the same as ground as far as an AC signal is concerned.

But when the cap is placed in parallel with the plate resistor, any noise in the HT is coupled directly to the stage's output.

I guess if your HT is pristine feeding the stage where the cap is used, it's not a big worry. This is most likely the case on the first gain stage in the preamp.

But when you're talking about the PI stage, there may be enough HT noise to make an audible difference when connected to ground vs HT.

Another one that can only really be solved though experimentation 

[HotBluePlates]

Regarding caps in parallel with plate resistors, whatever stage they're used on, isn't it better to connect the cap between the tube's plate and ground, vs the plate and HT?
...
But when the cap is placed in parallel with the plate resistor, any noise in the HT is coupled directly to the stage's output. ...

This is good thinking! But it will help you to change your perception of a capacitor...

Don't look at a cap as a "pipe for signal".  Think in terms of a frequency-dependent impedance: very low impedance (or "resistance") for highs, and a high impedance for low frequencies.

If there is noise on the B+ supply, it is likely to be 120Hz plus any harmonics (120Hz due to full-wave rectification).  The upper harmonics of that noise should be substantially reduced by the filter cap(s).

What reactance (the "a.c. part" of impedance) does a typical bypass, say 200pF, cap have?  X_c = 1/[2*Pi*120Hz*(200*10^-12F)] = ~6.6MΩ.

The cleanest d.c. in most amps is that which feeds the input stage. But let's say there is some 120Hz noise on it... The noise will then see a voltage divider made of that 200pF plate bypass cap and the grid resistor for the following stage (we'll ignore any additional voltage division due to the coupling cap).

So the 6.6MΩ against a 1MΩ (typical) grid resistor will reduce the noise by a factor of ~7.  And for most amps the 120Hz hum on the B+ near the first preamp stage is only a couple-millivolts.  Meanwhile the guitar signal will be several volts to tens-of-volts.

So you're probably looking at relatively little added noise, if it's even perceptible.  Some hi-fi designs really go crazy about trying to minimize power supply noise, but it's not always clear the effort is worth it, or whether they put so much effort into hum neutralization to allow for more slop in the basic power supply.

[Ken Moon]

Thanks for that well-stated reply 

Sometimes it's easy to obsess over little things, and it takes wisdom and experience to keep in mind the relative magnitude of things.

I built an amp with parallel input triodes to raise the SNR, but when looking on the 'scope at the results, I saw that the noise reduction from the parallel triodes was totally swamped by the noise generated by the 10k grid stopper. Oh well...

I'm always impressed by those, like you, who can see the forest AND the trees. I'm getting there, but at 61 yrs old, I better hurry up 

[PRR]

> noise reduction from the parallel triodes was totally swamped by the noise generated by the 10k grid stopper

Cathodes are around 3K hiss resistance. Two (1.5K) makes little difference against a many-K series grid resistor.

The elephant in the room (or cord) is the 10K to 100+K (at resonance) resistance in the guitar.

[2deaf]

Cathodes are around 3K hiss resistance. Two (1.5K) makes little difference against a many-K series grid resistor.

The elephant in the room (or cord) is the 10K to 100+K (at resonance) resistance in the guitar.

Could you rephrase that in the form of a quatrain?

[PG000]

My hum problem is now gone and the hiss got very low. Thank you for all of your help

[trobbins]

Could you indicate what you did?

[Ken Moon]

As a side note, I was looking at some Rivera/Fender schematics, and they had plate resistor bypass caps on several stages, but they were connected between the tube plate and the cathode, which allows you to put the cap right across the tube socket. (All of the cathode resistors were fully bypassed, providing the AC path to ground.)

So ease of layout may be the determining factor on where to put these caps.

[2deaf]

. . . So ease of layout may be the determining factor on where to put these caps.

They may have put them on the sockets to make the circuit more stable.  That's why I mount the cap. from plate to grid right there on the socket.  You're always just about one phase shift away from having a new, high frequency tremolo driver design and running some wire and having the cap. at a remote location just might do it.

[MFowler]

Having built this circuit many times I just use lower gain preamp tubes in V1 or V2 or both positions.  We also tube roll many times to find the best for each position.