recordingvalve preamp

[mozwell]

Dear fellow MEF people
I am an EE, with much more solid state than valve experience, & yes i know how to work safely with HV
I have managed to help some people here, whether it be faultfinding, or schematics, now i am asking if the kind people here with much more valve experience than me can go over my muddled thinking

I built this up & it sounds great, i was using it to record guitar straight into the audio interface.
I have very large heat sinks on the +/-12.6V supplies, these don't get hot at all.

When i first powered this up, it was quiet as a mouse, but after several hours being left on, the unit developed a small hum, which i narrowed down to the first gain pot (when i pushed slightly on the shaft, the hum was gone) so i replaced the pot
I also grounded out all the pot bodies.

My layout is star grounded with point to point wiring, the 0V for the valve & solid state output only tie together at one point, shielded wire from the input sockets direct to the valves.
On C5 (location D2) i get 198VDC & 3mVAC ripple, C29 (location B6) i get 225VDC & 6mVAC ripple, C28 (location B6) i get 235VDC & 20mVAC ripple (im just going from memory on these voltages)
ALL controls affect the hiss, if i turn down GAIN the hiss is gone, if i turn down the MASTER the hiss is gone, tone controls affect the hiss, the bright switch affects the hiss, so it looks like the hiss is coming from before all the controls, so around the first valve stage. All this is with nothing plugged in & the input sockets shorted to 0V at the socket by the sockets shorting contacts.
Input & output sockets are isolated from the chassis
All capacitors are WIMA types, except for the ones noted below
All resistors are metal film, except for the 1W types which are 5% carbon

What i am thinking, it is one of the following that is my problem, as i have a hiss not a hum
1. anode resistors are carbon film 1W types, i will replace these with metal film
2. the first valve itself may be noisy, so i will get another to replace it. I currently have an electro harmonix 12AX7WA in there
3. add some 100n 400V caps across the power supply filter caps (note i am using 47u 450V caps not the 15u caps on the schematic)
4. C6 & C9 are ceramic disc types, i will change these to orange drop or similar caps
5. C3 is a 47P 1kV ceramic disc, i will change this to an orange drop or similar cap

An ideas ?

Thanks in advance
Mozwell

[Leo_Gnardo]

No muddle, looks like a mighty nice preamp! Re # 4 & 5, there are no orange drop caps with those ratings. A good ceramic cap is perfectly fine. C6 has the effect of tipping up the frequency response & that will bring out hiss. Try doing without it, or you could put a multi position switch there to select "brightening" factor, realizing larger cap values will bring along more hiss. It's much like a "bright" switch on a standard Fender amp.

I find EH to be bright & kind of hissy, as they age get downright noisy. Last couple years I've come to like the 12AX7 JJ's been turning out, not the fancy ones, just ordinary. Compared to others, there's a good "fade to black" between notes. Maybe not everybody's taste, but they're inexpensive & certainly worth a try.

Metal film anode resistors, definitely the lowest noise.

Another noise source is the grid-stopper. You have 68K. Lower values will yield lower noise. Although carbon comp resistors are noted for potentially being noisy, their noise is proportional to voltage across the resistor, and there's none in the grid stop position. CC's also suggested as grid stops because they are very linear with respect to frequency (note radio gear is often built with CC for that reason). Consider MF & CF are spirals of conductor, look like inductors at high frequencies. You may be able to use 10K or less grid stop & reduce noise that way.

[mozwell]

Thanks Leo
I appreciate the comments & suggestions

It really did sound nice for recording, especially with the cabinet emulator circuit in, and when i add a cabsim vst & quad box IR file, its hard to tell from a real amp
The circuit went from really nice "chimey" cleans, to searing overdrive, all very controllable.

The main circuit is a JCM800 preamp, and the cab sim (i think) was a run off groove thing emulating the frequency response of a Marshall quad.
I know i don't get the overdrive characteristic of the phase inverter & power amp, or the dynamics of an actual speaker, but this gave me pretty good sounds (until the hiss) and it sounded a lot better than many of the vst amp sim plug ins available.

I will report back with what got rid of the hiss over the next few weeks when i get a chance to look over this thing

[loudthud]

Star grounding can be wonderful in theory, but in practice it usually falls short. In particular the spiky charging currents in the power supply can enter the signal path at the star due to it's imperfect nature. It's best to isolate these currents upstream from the star.

How is your star implemented?

What is the waveform of the hum? Is it 50Hz, 100Hz or a combination?

[MarkusBass]

I'm also not sure about the star grounding. Have you checked how it was done in JCM800? When I build such preamps, the input stage is usually grounded at the input jack (to the enclosure) - not to the power supply. Apart from this, I see that you changed several components: you have 2.2M on the input (instead of 1M), 82k in the cathode follower (instead of 100k) and the heater power supply is unusual (two valves in series). Have you verified that you have 12.6V on each valve?
It seems to me that such a supply is far from perfect because you may have 10V on one valve and 15V on the other. Also heater power supply elevation is something that should be considered. I would check what voltage there is on the cathode of the cathode follower and verify whether this is within acceptable voltages for 12AX7. Not every 12AX7 can stand e.g. 180V between cathode and the heater.

Mark

[g1]

What i am thinking, it is one of the following that is my problem, as i have a hiss not a hum

Is the current issue still the hiss, or a hum? I was under the impression the hum is cured and you are now chasing a hiss.
And if this is correct, was the hiss there initially or not?
If it is hiss, and it was not initially there, then the grounding method is probably not an issue.

[mozwell]

Thanks guys for your comments
When I built this it originally was noise free, so I don't think the star ground is an issue, I just noted it so as to provide as much info as possible.
The hum appeared after I left the unit on for 5 or so hours.
Good point about uneven sharing of heater voltage, I will definitely check that & if a problem, connect the centre of the heater string to 0V.
I have a 10 ohm & 47n from the ground of the input jack to chassis / mains earth. That is my only connection to earth.

I think the hum is gone after I replaced the gain pot. Now all I hear is the hiss, but until I get rid of the hiss I cant see if the hum is still there. I think the hiss is from the first valve, as all controls affect the hiss so the hiss has to emanate from before the gain pot

I will measure the voltages on the heater string
I will measure DC and AC voltages on HV power supply nodes & also on the +/- 12.6VDC supply
I will replace the anode resistors with metal film
I will add 100n to the HV power supply nodes
I will change the input grid stopper to 10k or less
Then I will pull the first preamp tube & see if the noise goes, ditto for the second preamp tube.
I will get some JJ's which are available near me & try them.

Thanks all for comments / suggestions

[mozwell]

Hi Guys
OK i managed to get some time to work on this last night
Voltages measured with my fluke 87
All readings with both valves in, bright switch on, nothing plugged into the input (inputs shorted through the input jacks) & all knobs on 11
Valves i used are MESA Russian 12AX7-A, with only about 20 hours use on them (the same valves from when the unit was hissy)

1. heater string, V2 pin 4 12.62VDC 1.1mVAC, V2 pin 5 0.15VDC, V1 pin 4 0.15VDC, V1 pin 5 -12.6VDC 1.1mVAC
2. power supply C28 234.5VDC 0.17VAC, C29 204.0VDC 1.4mVAC, C5 196.3VDC 1.2mVAC
3. V1 pin 1 145VDC 2.5mVAC, V1 pin 6 169.6VDC 4.8mVAC, V2 pin 1 118.4VDC 24mVAC, V2 pin 8 118.9VDC 24mVAC
4. Output measured right at the socket 0.0VDC 9.0mVAC this drops to 4.1mVAC with the bright switch off
5. V1 & V2 removed, output 0.6mVAC, V2 installed output 0.6mVAC, V1 installed output 9.0mVAC

So i changed V1 grid stopper from 68K to 1k, i changed all the anode resistors to metal film i added a 100n 630V cap in parallel to C5 the main DC filter cap for the first valve stage
V1 & V2 removed, output 0.6mVAC, V2 installed output 0.6mVAC, V1 installed output 0.6mVAC

So i think it was just a noisy anode resistor on V1
Voltages on my heaters are ok for each valve & the plate to cathode voltage on V2B looks well within datasheet max ratings
I will get a spare JJ 12AX7 ECC83S "just in case" one of the MESA tubes goes south

I will check it out connected to my recording interface over the next few days but i think its fixed (fingers crossed)
Thanks to all who have helped me with my problem, I will continue to post schematics or provide suggestions to help others here.
This is a great place, (mostly) kind & helpful (apart from those who dont like crate).....

Cheers
mozwell