Let's not forget TV damping tubes like PY500A and the likes. Great as rectifiers, low internal resistance, high current.
I'm using a version of what Eli Duttman describes as a "cockeyed bridge" on my current breadboard project.
The UF4007 is a faster / quieter step up from the 1N variety and still very cheap. My understanding is that the snubber caps keep any noise that the diodes do produce out of the other transformer windings.
The 6AX4GTB is super cheap, provides a very slow warmup and, like other tube rectifiers, has no switching noise. It's good for 165mA and its 6.3v heater only draws 1.2A. There are other damper diodes that can handle more current or I suppose you could use a pair of 6AX4s.
Last edited:
Nostalgia is the only reason I can think of. I keep using 1N4007 and always model the PI filter in PSUD2. I've never had issues with switching noise or other ill effects that would be resolved by upgrading to UF4007. I should add that I work with line level HiFi and the occasional guitar amp mostly.
Last edited:
I can't say that I've ever heard diode switching noise but then I've never done any tests to listen for it. And I don't have the equipment needed to measure it. The price difference between 1N and UF parts is negligible so I figure I might as well use the UFs.
The reason I like using (indirectly heated) tube rectifiers is because of the slow startup. You can use a thermistor but they don't come anywhere close to an indirectly heated tube. Even a directly heated rectifier is probably a bit more gradual. Nothing wrong with nostalgia, either.
And many guitarists seem to like tube rectification because they consider the "rectifier sag" to be part of their "sound".
That said, I've got numerous vintage amps that use SS rectification and they seem to have survived the faster startups. The "cockeyed bridge" I posted earlier seems to offer a bit of both worlds.
The reason I like using (indirectly heated) tube rectifiers is because of the slow startup. You can use a thermistor but they don't come anywhere close to an indirectly heated tube. Even a directly heated rectifier is probably a bit more gradual. Nothing wrong with nostalgia, either.
And many guitarists seem to like tube rectification because they consider the "rectifier sag" to be part of their "sound".
That said, I've got numerous vintage amps that use SS rectification and they seem to have survived the faster startups. The "cockeyed bridge" I posted earlier seems to offer a bit of both worlds.
I like your [Eli Duttman]'s 'cockeyed bridge' 🙂 Thanks for that.
Couple things of note (IMHO):
Member Mark Johnson has contributed a great project, "Quasi-modo the bell ringer", and "Cheapo-modo . .", or some such. Building the project may not be worth it to you, but please read a few pages of the lengthy coverage (it's something like 7 or 8 years old now, and still kicking . . 😱 ). The technique can largely suppress / eliminate EMI from SS rectifiers.
And you don't need valve rectifiers to provide 'rectifier sag' -- a resistor or resistor+PTC thermistor can provide that.
Best Regards
Couple things of note (IMHO):
- TV Damper diodes often have enticing ratings, but the duty cycle for their service is typically below 10%
- eliminating the noise produced by silicon rectifiers requires adding a resistor to dampen the rectifiers' commutation energy
Member Mark Johnson has contributed a great project, "Quasi-modo the bell ringer", and "Cheapo-modo . .", or some such. Building the project may not be worth it to you, but please read a few pages of the lengthy coverage (it's something like 7 or 8 years old now, and still kicking . . 😱 ). The technique can largely suppress / eliminate EMI from SS rectifiers.
And you don't need valve rectifiers to provide 'rectifier sag' -- a resistor or resistor+PTC thermistor can provide that.
Best Regards
High vacuum rectifiers don't really eliminate rectifier switching noise; that's caused by the rectifier doing its job switching currents from the transformer's secondary. What's different is the very small (and almost constant) rectifier capacitance. The smaller capacitance pushes the di/dt turn-off ringing of the transformer's secondary inductance up to higher frequencies, and also doesn't couple it as much downstream through the B+ supply. Other coupling still happens, so damping RC across the secondary is still useful.
Mark Johnson's work is very important, but for DIY folks a brute force (overdamped instead of critically damped) RC choice does the job. For the usual 5something4 rectifier situation 1000pF and 1000 Ohms (give or take a lot) works.
All good fortune,
Chris
Mark Johnson's work is very important, but for DIY folks a brute force (overdamped instead of critically damped) RC choice does the job. For the usual 5something4 rectifier situation 1000pF and 1000 Ohms (give or take a lot) works.
All good fortune,
Chris
Thanks Chris,
Your 1000pF / 1000 Ohms should give valve lovers a good general solution.
As long as everyone realizes than <more capacitance> doesn't mean <more damping> -- that it is resistance that does the damping. (I struggled with that once upon a time . . 😕 )
And lower-capacitance / higher-frequency usually means Higher Radiated EMI, even if Conducted is lower. 😉
Best Regards
Your 1000pF / 1000 Ohms should give valve lovers a good general solution.
As long as everyone realizes than <more capacitance> doesn't mean <more damping> -- that it is resistance that does the damping. (I struggled with that once upon a time . . 😕 )
And lower-capacitance / higher-frequency usually means Higher Radiated EMI, even if Conducted is lower. 😉
Best Regards
Is the b+ is out from the centre tap provided by 6ax4?
For RC filter, isnt 1000pf abit small? Or what am i missing?
Yeah, the B+ comes off the cathode of the 6AX4GTB which is connected to the first PS cap, just as it is with a more typical design.
A bit of background . . . Eli originally proposed using this circuit and some variations when people were first getting interested in using Cree Silicon Carbide Schottky diodes, higher voltage versions of which had only recently been introduced. The Schottkys have zero switching noise but they were quite expensive at the time. As a result, some builders were hesitant to use them. I believe they are cheaper now.
Eli proposed a lower cost solution which employed much cheaper UF diodes, which are supposedly lower noise than the older 1N types but still very cheap. When snubbed, whatever noise they have is mostly eliminated. His original design used a single Cree Schottky in the position of the 6AX4GTB. Someone asked about slow startup and he suggested the version that I'm using. Any indirectly heated rectifier in this position will offer slow startup, just make sure it's specs are sufficient otherwise.
He used to post here and on AA fairly regularly but he hasn't been active lately. If you search you'll likely find more details.
I have limited experience with damper diodes but I know others have used them for many years, including in some very high end, expensive amps. See Thomas Mayer's designs, for example. I'm not aware of any reports about them having shorter lifespans. Perhaps others will comment.
I'm not particularly knowledgeable about the the technical details, I'm pretty much just a copy and paste builder. As I understand it, the snubber caps are designed to reduce switching noise but I'm unable to offer any analysis beyond that. Eli Duttman is pretty sharp, though, so I trust that it's well designed. I've other designs that use snubbers and I've used them myself in a hybrid (Graetz bridge) PS. I've never seen resistors used but perhaps I've just missed it. Please post a schematic, if you can.
As I said, I've never been able to hear diode noise and I have no way to measure it. I will often employ parts and designs that are considered technically superior, even if I can't hear any difference, as long as they aren't complex or particularly expensive.
The 1000 ohm / 1000pF bit was about the snubber to dampen the power transformer secondary's leakage inductance.
I think you're thinking about what I would call bulk filtration -- and yes, 1000pF would be a complete waste there.
The Quasi-modo threads offer tons of insight into the point of it all, without requiring EE skills to perform the analysis.
Just my non-EE 2¢, but I'm pretty sure a snubber isn't a snubber without a resistor to dissipate the energy.
Cheers
Thanks for an excellent point. Mark Johnson's general recommendation, made in the context of semi-con rectifiers in lower voltage supplies, is to parallel the snubber RC with a capacitor a coupla times larger than the C of the snubber. Something along that line might be included in higher voltage HV rectifier supplies.
All good fortune,
Chris
Sorry for my ignorant. Try to search Quasi-modo thead but to no avail. Please kindly attached the schematic
Sorry -- I spelled it wrong! 😕 But I think the concept may be more useful than the schematic/project. Here it is:
Mark Johnson's QuasiModo test bit
Cheers
Mark Johnson's QuasiModo test bit
Cheers
Generative references are
Morgan Jones "Rectifier snubbing - background and Best Practices", Linear Audio vol.5 p.7
Mark Johnson "Soft Recovery Diodes Lower Transformer Ringing by 10-20%", Linear Audio vol.10 p.97
All good fortune,
Chris
Morgan Jones "Rectifier snubbing - background and Best Practices", Linear Audio vol.5 p.7
Mark Johnson "Soft Recovery Diodes Lower Transformer Ringing by 10-20%", Linear Audio vol.10 p.97
All good fortune,
Chris