• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Help - noisy phono stage

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

I have just wired up my new phono stage and am being plagued by noise. Hiss to be exact (no hum at all). The hiss is present at all times, when the cartridge is connected (MC) via the phono leads from the turntable and also when the input is shorted.

The hiss was also present prior to installing the MC step up transformers when I was doing basic testing with a cheap MM cartridge, however at that point it was a little worse.

Maybe advisedly I have built the phono stage into a single chassis (PSU and all) due to space constraints, maybe this is something that I will just have to address and put the PSU into a separate enclosure but I would like to avoid that if possible (however if that is the only way then so be it).

The circuit is a modified version of dsavitsk phone which can be found here or here with the 1st tube replaced with a triode wired E810F running with at 25mA currently biased at 1.4V (56R in the cathode) which gives an anode voltage of ~125V. The 51K resistor after the first 6GK5 has been changed to 56K due to the different anode resistance of the triode strapped E810F and the zener setting the anode voltage of the 6T4 in the output position is set to 116V (100V + 16V) so that I am dropping a little less voltage across this CCS at 20mA.

The HV regulator is a kit based reg from janneman using DN2540s and there is a standard LM317 regulator for 6.3V for each channel. The 6.3V is raised up 30V above ground. The HV looks like it only has a handful of mV ripple on its output using my DVM.

Ground is raised via a combination of 0.1uF, 10R and a pair of diodes.

The ground from the HV regulator goes direct to the raised ground point, as does the ground of the output jacks. There is then a ground bus that runs from input jacks, through the 1st, 2nd and 3rd stages before being connected to the raised ground point. The negative terminal of the input jacks are connected to the chassis via 10nF caps.

The CCS's themselves are DN2540s cascaded with LM317s and have grid stoppers which are connected to the ground bus via 100nF film caps. The positive ends are also connected to the ground bus via 470nF film caps (except for the CCS in the cathode of the 6T4).

There is shielding between the PSU and the amp section and the perforated sheet that the amp section is mounted on is isolated from the chassis due to the way that it is mounted but it has a ground lead also (to chassis).

Attached is a picture of the amp section and then hopefully in the next post the PSU.

Using my DVM I measure around 10mA AC on the output with the input shorted, however saying that the phono stage seems to have a lot of gain - much more than my last phono stage which is something that I will need to look into/attenuate somewhere.

What can I try to resolve the hiss without separating the PSU into a 2nd box?

Is the grounding scheme OK or should I look to do local star ground nodes all connecting back to the raised ground?

Should the film caps to remove noise from the CCS's be connected to the signal ground or should they be grounded separately - to the chassis or raised ground?

Thanks in advance for any help - Ian
 

Attachments

  • image022s.jpg
    image022s.jpg
    91 KB · Views: 428
First, I don't think that the problem is the power supply proximity. The supply is much more likely to cause inductive hum issues. This would affect the MC transformers most, unless they are well shielded. Since you don't hear hum, I wouldn't look there yet.

I can see three initial possibilities: Tube noise, oscillations (tube or CCS), or noisy zener. I would tend to put the zener low in priority, since it is on the follower stage, and not amplified.

That leaves tube noise or oscillations. Have you tried different tubes, especially in the input position? If switching out tubes doesn't change the level, then they are probably not to blame.

So that leaves oscillation. I'd suspect your CCS. Try a simple DN250, without the cascade, on one channel and see what happens. Maybe the speed of the DN2540 doesn't match well with the voltage reg.

Sheldon
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Additionally to Sheldon's sound advices, if things don't go better, I would use resistors as anode loads for a quick experiment. That would lower the gain, and would show if the hiss comes from CCS loads, something that I have experienced before with enough mA active anode loading. Especially with the Jung Depletion-317 cascode combo. I would also check my HV for wideband noise, if the noise won't go away on first experiment.
 
OK, many thanks.

I think that gives me some options to try out to try to resolve issues, I think my order due to simplicity will have to be:

1. Swap out the E810Fs
2. Resistive anode load for one channel E810F - see if that makes that channel quieter
3. Bypass/unhook the regulator and replace with a simple CLC
4. Swap one of the CCSs on one E810F for a simple DN2540 one

I also wonder whether I could have instability in the last stage with the CCS on top and bottom of the 6T4; could this give me ~10mV noise? Theoretically I could check for noise on the input of the 6T4 as it is less than unity gain and if there is none then this could be the problem - or am I barking up the wrong tree?

If it turns out that the CCSs are the problem what would people suggest as a replacement - ideally using the DN2540s. There is at least 20V to play with (could be more if I change the biasing slightly).

Thanks, Ian
 
ianc13 said:
I think that gives me some options to try out to try to resolve issues, I think my order due to simplicity will have to be:

1. Swap out the E810Fs
2. Resistive anode load for one channel E810F - see if that makes that channel quieter
3. Bypass/unhook the regulator and replace with a simple CLC
4. Swap one of the CCSs on one E810F for a simple DN2540 one

Sound plan.

ianc13 said:
I
I also wonder whether I could have instability in the last stage with the CCS on top and bottom of the 6T4; could this give me ~10mV noise? Theoretically I could check for noise on the input of the 6T4 as it is less than unity gain and if there is none then this could be the problem - or am I barking up the wrong tree?

Makes sense to try this - isolate a stage at a time. Easy to do. Always check the easy stuff first.

ianc13 said:
[BIf it turns out that the CCSs are the problem what would people suggest as a replacement - ideally using the DN2540s. There is at least 20V to play with (could be more if I change the biasing slightly).
[/B]

A single DN2540 has pretty good performance. I wonder how many people can actually hear a difference between that and a cascaded pair. But people have use a cascaded pair successfully. Probably no harm in increasing the gate stopper resistor (assume you have something like 1k) to 2k or higher.

Sheldon
 
Hi,

New tubes in the 1st stage did not make any difference - still 10mV noise on the outputs with the inputs shorted and there is ~10mV on noise going into the final stage so it isn't that!

My gate stopper is currently only 100R CC :smash: - nowhere near 2K or anything like that so maybe that could be allowing the CCS to oscillate? Time to order some ~2K CC's I think, going to need them even if I go to a single DN2540 as opposed to a cacscode.

Thx Ian
 
ianc13 said:
New tubes in the 1st stage did not make any difference - still 10mV noise on the outputs with the inputs shorted and there is ~10mV on noise going into the final stage so it isn't that!

My gate stopper is currently only 100R CC :smash: - nowhere near 2K or anything like that so maybe that could be allowing the CCS to oscillate? Time to order some ~2K CC's I think, going to need them even if I go to a single DN2540 as opposed to a cacscode.

Good progress.

Yes, stoppers for this are usually 1k, so I'd go at least that high.

Sheldon
 
ianc13 said:
The circuit is a modified version of dsavitsk phone which can be found here or here with the 1st tube replaced with a triode wired E810F running with at 25mA currently biased at 1.4V (56R in the cathode) which gives an anode voltage of ~125V.

Hey, you built my circuit :) Mine is very quiet. I do have the power supply a few feet away, though I agree that this is not likely your problem.

I'd try working from the back to the front. Listen to just the 6T4 output stage with the grids grounded and everything else disconnected. If that is quiet, then add in the previous stage. Do this one at a time until you find the noisy stage. As mentioned, a CCS or those beastly E810F's would seem the likely culprit. The CCSes, if they are doing their jobs, should block PS noise before them.

While the zener shouldn't inject noise, there is some argument that it is a good idea to bypass it with a big cap, or use some fancier shunt element. I have not played with this at all, so I can't give any concrete suggestions.

-d
 
I don't think the noise is coming from the last stage - well not the majority of it anyway!

It is on both channels though and is of similar intensity.

The PSU seems to present less noise than what I get at the outputs (~5mV against ~10mV at the output phono jack).

I think my first port of call now is going to be to replace the first CCS with a resistive load on one channel for an E810F and see if that makes any significant difference to that channel. Then stretch that to the 2nd stage on the same channel to see

If so then it is the CCS's that are the problem (or their interaction with the regulated PSU), going to be a couple of days before I can get my hands on any 1K and up carbon composites so if it is the CCS's then I can at least try a bigger gate stopper then.

Thanks for all your help so far. Ian
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.