A while back, I built the RJM phono preamp. The result can be seen here. It's been working great, but I started noticing an externally audible ringing sound a while back. As in, nothing coming through the speakers, it's actually coming from the amplifier itself. I troubleshat it, and couldn't find anything obviously wrong. From memory, I reflowed all the solder joints in the power supply section, as it was all I could really think to do at the time. Admittedly, I'm still fairly new at this sort of thing, so the odds that I missed something, even obvious, are high. It finally got bad enough that I could track it down, and it's coming from the RECTIFIER tube! Whatever was ringing got bad enough and loose enough inside that I was able to tap the tube to make it stop. I figured it was a faulty tube, and swapped in a new one. It's doing the same thing (though not as loose yet).
I took a video:
Camera is close so you can easily hear the noise coming from the tube. Toward the end, you hear me tap the tube, and the sound stops. Really, I think it disrupts whatever is resonating, and it has to build back up to be audible again.
Any suggestions for what might be causing the ringing would be appreciated. I'm stumped.
I took a video:
Camera is close so you can easily hear the noise coming from the tube. Toward the end, you hear me tap the tube, and the sound stops. Really, I think it disrupts whatever is resonating, and it has to build back up to be audible again.
Any suggestions for what might be causing the ringing would be appreciated. I'm stumped.
Leaking magnetic field from the power transformer could set the plate ( usually steel) in motion.
Locate the rectifier further away
Locate the rectifier further away
Whats the size of your first filter cap.if its to large the repetative peak plate current will be to high stressing the rectifier tube. For exAmple a 5u4 is rated for only 30 to 40 uf on the first cap input filter.
Location of the transformer is 4" away with a large choke in between the two. Would choke cause the same issue?
Is there a quick way to shield the choke as a test of this theory? Grounded steel plate or something comes to mind, but want to be sure I understand that correctly.
Short the choke. Perhaps you will get a little hum in your audio, so turn the volume down, but at least you cannot hear the choke any more.
Reagards, Gerrit
Reagards, Gerrit
Don't overthink this.
Change the EZ81 if it worries you.
Temporarilly short the choke to stop current flow in the choke for test purposes.
The choke and or mains transformer is not likely the cause.
Change the EZ81 if it worries you.
Temporarilly short the choke to stop current flow in the choke for test purposes.
The choke and or mains transformer is not likely the cause.
I changed the EZ81, and it's doing the same thing, just not as bad (yet). Memory is not my strong suit, but I'm pretty sure the reason I had a spare was because of this exact issue and that exact attempt at a fix.
Did you use the choke-input power supply? Some chokes and rectifiers don't like that. Too much AC ripple and they will hum like that. Try a .47uF cap to ground before the first choke. It won't change the B+ but it might get rid of the buzzing.
It's exactly as on the schematic linked in the first post with the LCLC power supply at the bottom of the page. Are you suggesting making it a CLCLC filter? Is that a thing?
What I'm suggesting is that choke-input supplies often require a special choke (called a "swinging" choke in the old days) that is designed to handle the very high AC ripple it will see from the rectifier tube. A choke that isn't designed to handle that AC ripple will sometimes buzz or "dance." I think this is what you might be experiencing here. The cure is to apply a small cap, usually .47uF, from the rectifier output to ground. This preserves the benefits of a choke-input supply, but also relieves some of the stress on the choke so that it (and/or the rectifier) doesn't vibrate from stress.
Yes, a CLCLC supply is a "thing." It's just a cap-input supply and choke followed by another cap and choke. There's nothing exotic about that. But in this case, we're keeping the benefits of a choke-input supply while giving that first choke a bit of a break.
Yes, a CLCLC supply is a "thing." It's just a cap-input supply and choke followed by another cap and choke. There's nothing exotic about that. But in this case, we're keeping the benefits of a choke-input supply while giving that first choke a bit of a break.
Another thing that's missing from that power supply is a bleeder resistor to ensure a minimum current draw on the rectifier. If the 6DJ8s should fail for any reason and cease to draw current, the B+ will skyrocket and possible take out some caps and other things. Every choke inpout supply should have a bleeder resistor to maintain a minimum current draw. I'm guessing that a 10 watt, 15K resistor in parallel with C1 would draw 15mA and that would probably suffice. But I am not an expert in this area. Maybe Chris Hornbeck or someone else could advise as to what would be the appropriate safety measure here. Or the supply should be modelled in PSUD to confirm that.
Actually, I'd suggest that C1, C2 in the power supply and C7, C8 bypasses all be rated somewhat higher than 400 VDC. 500 VDC rating is easily available and above the expected peak if unloaded or at turn-on. That peak is sgrt of 2 times unloaded transformer voltage and assume a lossless rectifier and slightly high line voltage, call it +450 VDC minimum safe rating.
Bleeder resistors to maintain "choke input" conditions can be estimated pretty well by setting the load in kOhms equal to the inductance of the input choke in Henries. That inductance will be higher than normal working value because it'll be lower on the B/H curve - good thing because it's very stringent (and wasteful). Does PSUD take iron core characteristics into account? Don't know, sorry. Probably easiest to just spec the caps higher and call it good.
I can't hear the noise on the YouTube video, so I'm not much help.
All good fortune,
Chris
Bleeder resistors to maintain "choke input" conditions can be estimated pretty well by setting the load in kOhms equal to the inductance of the input choke in Henries. That inductance will be higher than normal working value because it'll be lower on the B/H curve - good thing because it's very stringent (and wasteful). Does PSUD take iron core characteristics into account? Don't know, sorry. Probably easiest to just spec the caps higher and call it good.
I can't hear the noise on the YouTube video, so I'm not much help.
All good fortune,
Chris