Thoughts of apparent noise on ground - due to power tx?
Hey all! Fleshing out a first wave of DIY designs related to microphone preamps, tube summing amps for a recording console, etc... Haven't had much to post, as you all dive deep into esoteric DIY-ing, but mainly use the knowledge base here for troubleshooting stumps...
Here's one that's giving me hell.
I seem to have, if I'm diagnosing correctly, about 1mV of (what looks like) transformer induced voltage on the aluminum anodized chassis I'm building a high gain mic pre on. (The mic pre is dual-pentode; 6SJ7 into EF86 into 6FQ7 CF, unbalanced outs at the moment).
I have observed via o-scope that this ~1mV or less noise signature is on:
- every part of the chassis and every ground in the amp
- the cathodes of the 6sj7 and ef86's
- the heavily filtered B+ line
I wouldn't normally complain about a small dab o' hum at very high gain, but this is a mic pre I'm building for ribbon mics, so it needs to be pretty damn quiet. I'm measuring this noise with the preamp *not* plugged into *anything else*, just terminating the inputs (150R on each leg on primary side of mic xfmr) and outputs (100k to ground after CF output cap). I've measured the same hum/noise levels with the preamp terminated this way, and actually terminated to a mic and mixer. This hum is not in relation to any other interconnected equipment, but is rather being self-induced somehow.
Here's my troubleshooting so far:
- Made sure anodizing is scraped off, making good contact around the chassis. I'm still suspect that this anodized chassis could be the issue. Read on...
- First, wasn't using a star ground scheme. Went through extensive pains implementing a good star ground. No changes.
- Transformer B+ sec CT and first filter cap grounded together to chassis; other two filter caps along with seperate stage grounds all star grounded to second chassis ground point. I have tried to change this to ground at the same place as the CT/first filter, no change in hum.
- Power transformer mounted on outside of chassis. No change if lams are grounded to chassis or not. Rotating PT makes no difference.
- O-scope plugged into same power strip to observe hum. Also tried different strip.
- Tried different grounding spots on the chassis. Disconnected parts of the chassis to try to reduce hum. No luck.
- Tried different power transformer and different voltages. Lower voltages (hence lower currents) slightly decreased this noise on this ground. This leads me to think of some kind of grounding current issue...
- Tried RC filtering as well as LC filtering. No change. It doesn't look like B+ ripple in the first place, anyways.
- Turning AC power off to power transformer stops the hum.
- With the PT 'on', I put a DPST switch to turn off the AC filament line. VERY slight change in sound of hum, but otherwise, no change.
- Tried using filament CT as well as dual matched 100R for filament ground sense. Nope. Tried humdinger balance pot. Nope.
- Tried various outlets in house. Nope. Tried it at work. No change. Injected noise and level is the same in a variety of houses/places/etc.
- Pulled various tubes, tried various tube types, makes, etc. No changes (other than slight gain differences due to tube strength/make)
- Shielded all AC-coupled cabling to/from stages. Cleaner waveform! No hum change.
- No noise change when (momentarily!) lifting the power ground from chassis.
- Noise reduced when I changed power cord ground to chassis from close to the PT to next to (but not the same connection as) the B+ CT / first filter cap ground point.
At this point, and at the start of it, of course, the design is questioned. I don't have the ability to scan anything, but it's a very simple design, two grounded cathode pentodes into a bootstrapped CF. I can isolate each stage and see the noise amplified in each, not dependent on the others. I feel very confident that the design is not to blame, nor is the layout. This is built on perfboard. The PSU part (diodes, first filter cap, etc) is on the wayy other side of the 2RU chassis as the rest of the stages. Lid on/off the amp doesn't change anything, either.
Since I have tried two different power transformers (and even different power cords!), updated grounding schemes to no difference, etc, my suspect is that either this chassis isn't grounding well (hard to believe - it's aluminum!), or that perhaps there is some kind of PT secondary imbalance that's causing there to be noise on the ground. All other tube equipment, amps, etc that I've built or rebuilt all have very quiet 'looking' ground and operate SILENT. However, this build on this chassis...
Any thoughts? I'll try to get a pic of where the build is at now for any troubleshooting in a few days (after new years). Am I neglecting something, here? My thoughts are that this slight noise on ground is getting into the grids and/or cathodes of each stage and then getting amplified. (The waveform is inverted at the plate in respect to what is observed on ground/cathode).
A basic question, too:
Can signal ground noise be coupled into the cathode of a grounded-cathode gain stage, both bypassed or unbypassed?
And, can signal ground noise creep into the grid via the grid leak resistor (or potentiometer)?
Thanks for any thoughts or insight. The amp sounds AMAZING despite this noise/hum.
Last hint: it looks similar to Fig. 2 here: Earthing (Grounding) Your Hi-Fi - Tricks and Techniques
Yes, I know that's a MAGNETIC induced waveform. Hence my dilemma. I'd say 'similar', as it's not symmetrical, but seems to look close to that 'M' shape, and (if my memory is right) it's happening at 60hz.
Use a battery powered O-scope :-)
I had a similar problem with a mixer with a gain of 200 for microphones.
A couple of millivolts amplified ground noise amplified by 200 is massive !
I tracked it down to resistance in the pcb tracks.
I shorted the pcb ground out with wires and this improved matters a lot.
Not sure if it was just cheap chinese pcb copper but I should really have added a ground plane, will know better next time.
I'm seeing this noise on the chassis itself, seemingly everywhere, as well as on every point of the star ground. So, I feel it's not an interconnect problem. On distant points of the chassis, however, the ground noise/hum is louder on the o'scope. This noise is also directly on the first filter cap/B+ center tap as previously mentioned, and this amp isn't pulling much current... perhaps 50ma peak on b+.
If you are seeing 1mV of noise everywhere you probe (but not on the signal path, as that has increasing noise levels due to amplification), then I suggest you are indeed 'looking' at noise from a loop that includes your O-scope earthing via mains.
Your original post was a little ambiguous about what 'noise' was being reduced (ie. amplified output noise, or O-scope observed level on say a non-signal path point).
You don't happen to have a switching power supply somewhere in the mix do you? Noise can come in on the ground from a switcher PS.
no switching power supply. classic full wave cetner tapped rectified with 4n1007 diodes. i've done both little and lots of filtering with no change. tried .1uf bypass caps on c
the noise, which i'd describe as a hum/buzz (so, 60hz with lots of harmonics), is the *same* throughout the original post, to clarify.
so, the o'scope puzzle, eh? here are my troubleshootings about *that*...
1) doesn't happen with any other electronics on my bench. period. never has. so, if it is something that is just 'seen' with the scope, then something in this chassis itself is causing a ground loop. :confused:
2) i've also done scope probes with and without the scope probe 'ground' clip on the chassis. all that does is clear up some 'fuzz' i see on any probed signal. you know, this is just kinda standard scope procedure. but, to note, i've worked the scope with ground clip to chassis (or star ground) to no difference (other than a little clarity on the waveform)
3) the 'look' of the waveform produced by the probe when hovered close to/near the power transformer is very similar to the noise that i'm seeing after amplification. when i'm using a large power transformer, i can clearly see about the same waveform on the chassis being conducted. when using a smaller voltage power transformer, i see a very small but similar waveform ripple on the chassis. my rationale, so far, is that since I'm doing more R-C stages and voltage knockdown with the bigger transformer, more current is being used and such the chassis current is higher.
scope is a tek 465, btw. i've done (and still do) great troubleshooting with that thing.
so, let's say, for argument's sake, that I don't have a scope, or that my thought that I'm seeing a 'noisy' ground is inaccurate.
my b+ is filtered very nicely. 68uf after rectifier, 20H choke + 100uf, 4.7k + 100uf (this feeds the 6fq7 CF's and ef86's), then 2.2k + 220uf (to feed 6sj7 input stages).
i have not tried running DC to filaments. i've relooked and re-twisted/tried new layouts for the AC twisted filament wires. the filaments go to the 6fq7 first, then each ef86, then each 6sj7, one at a time, in parallel, of course. however, i've read and deduced that filament hum would be seen as (mostly) pure 60hz, no? and not quite this mangled hum/buzz i'm experiencing. would anyone suggest i try DC for filaments? it's one of the only things i haven't tried. however, my test i noted to turn off the filament line to the tubes (as b+ is still active) for just a few seconds while the tubes are still conducting yielded no reduction to the hum/noise, so this led me to feel that a filament AC problem is not what's causing this.
thanks for the thought help, everyone, so far. i'd like to solve this and add it to the knowledge database here!
i'll try to get some pics in the next few days. i'm moving cross country on friday...
20H choke sounds pretty substantial. The pics would be very helpful. I can't imagine the heater wiring twist being an issue if tightly twisted and layed against the ground plane.
What waveform do you get with the probe ground and tip both connected to the same point on the amp chassis, and other normally probed spots?
It's always worth having a 6V battery available for troubleshooting and preparing high gain amps. Two quick configurations that can help with a better appreciation are (a) just powering the input tube(s) with 6VDC, but leaving the 6VAC wiring pretty closeby, and (b) powering the whole heater chain with 6VDC, unconnected at the transformer end.
Have you tuned the humdinger pot for min hum? Was it very sensitive to pot wiper position?
|All times are GMT. The time now is 01:35 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Copyright ©1999-2016 diyAudio