[sdiy] sources of hum and hiss in a solid state guitar amp (LM1875) (slightly OT)

[anthony]

I've been so impressed with the performance of the tube amp that I just made 
that I am now completely unsatisfied with the performance of my Epiphone 
EP800R solid state 15 watt amp. The hiss and hum on it are atrocious. But 
since it has a cheesy but super-cool spring reverb I've been willing to 
overlook this. I'm sort of at a loss on where to start.

I do not have any sort of empirical baseline performance experience data 
because when I got this amp at a Pawn Shop it was a total basket case. The 
PC board and one of the pots was broken. I did a really good job fixing the 
board, but I only had a 100kOhm pot that would fit where the 250kOhm pot 
that broke was  - that may be a small part of the problem. What happens is 
there's a lot of hiss at high gains and at various points along the the 
travel of the gain and volume controls. Added to the mix is what sounds like 
some garbled oscillations at various points. This also is prominent when the 
reverb control is turned way up (which is really all the time since lower 
levels are virtually inaudible). What it ends up happening is a sound much 
like what you get from a BBD device with the lock running really slow. But 
of course there are no BBD's anywhere inside.

The preamp looks like a Tubescreamer except the clipping diodes are 2 red 
LED's. I changed these to 4 red LED's and 4 Ge diodes - symmetrically. This 
mod didn't change the hiss and hum (as would probably be expected). The 
original op-amp was an NJM4558  (the newer ones with with a 5-digit date 
code). I switched this to an NJM5532DD, but this actually sounded a lot 
worse. (The 5532 was also a newer 5-digit date code unit - I don't think any 
of their newer parts are up to snuff.) I replaced this with an older 
JRC4558DD whcih was a little better and I liked the tone from it a lot so I 
kept it in. I don't think I like the way LED's clip so I think I'll be 
making a small board with around 10 Ge diodes on to fine tune asymmetric 
soft clipping while retaining a relatively high level of clean headroom. 
That was really the first thing I didn't like about this amp when I got it 
going: it clipped way to early. Since I already have a rad distortion pedal 
I don't need to pretend this is a tube amp (OK so maybe giving it 
asymetrical soft-clipping is pretending it's a tube amp, but oh well...). I 
only used LED's because it was easy to get a lot of headroom with few 
diodes. Too bad it sounds like crap... I think I'd get great tone with a 
bunch of Ge diodes with only a difference in one or two diodes. I've tried 
experiments using Si in combo with Ge and concluded that all Ge diodes sound 
better (but this was in a pedal that used diodes to shunt the signal to 
ground).

The power amp is an LM1875. I read a little about these and was surprised to 
see people using them in sort-of hifi DIY audio. Actually I guess the 
preferred chip is the LM3875. Anyway I had thought of either swapping in an 
LM3875 or increasing the power supply voltage. It looks like this amp is 
using  a PS voltage a lot smaller than the max rated PS voltage. I'd 
actually prefer to not have to buy any new specialty chips - especially if 
I'm not even sure I'll like the result.

I'm really having a problem justifying keeping it as a solid state amp - 
especially given my early success getting a great-sounding amp. It has this 
cool tweed covered box and just about the perfect number of controls for 
what I'd want on a tube amp. (The only thing lacking would be tremolo 
controls, but reverb has priority over tremolo anyways.) There's a schematic 
for a perfect amp with tons of gain and reverb and everything on the AX84 
site. But I think - since I have enough Mullard EL84's - I want to do a 
push-pull amp: something like an AC15 or a GA15 with reverb. I'll still have 
1 EL84 left over to make a single-ended amp (or mod an Epiphone Valve 
Junior - which is most likely what I'll do with it...) I'm not sure if the 
reverb spring in this amp will work with a tube amp. But that's not really a 
big deal since I want a "surfier" spring anyway.

The big problem with making a push-pull tube amp in this enclosure is that 
the enclosure is a lot smaller than I'd like to mess with, so if anything 
I'd probably just end up making a simple single-ended 7W amp and save the 
pair for a 12" combo.

But in reality that's not really a simple solution to the problem at hand. 
What I need to do is figure out where the hiss and hum are coming from and 
do things to fix those. I had thought of putting an LM380 amp right before 
the reverb spring because the reverb on this amp is wimpy. That made me 
ponder how maybe at least PART of my problem is impedence matching. I 
noticed that the clock hiss from my DOD flanger/doubler is horrible with my 
solid state amp (even way down the chain), but barely perceptible with the 
tube amp (and sounding way sweeter with the tube amp). I guess it's obvious 
that the major source of hum to eliminate is the power supply. I have some 
good caps to put in so that's easy. The jacks are isolated from the chassis 
so no worries there.

What would be the best way to cut down on hiss and oscillation while 
maintaining the best possible tone for a guitar amp? It just now occured to 
me that maybe the LM1875 is a source of hiss and maybe reducing its gain 
(and making up for it in the preamp since I actually modded it for more 
headroom anyway) would cut down on the hiss a lot.

I could - even using the existing power supply - make a preamp using a 
12AU7. I have some sweet RCA side-getters that would be perfect. Getting a 
quite tube preamp seems more staightforward to me at this time. Maybe I'm 
missing something.