I've built plenty of transistor and IC amps in the past, and so for this foray into the hot world of valves I intended to be purist about it all.
That purist attitude quickly goes out of the window when it comes to rectifiers. I use silicon and if I feel the need for sag, then the bridge diodes get 150 ohms in series. I'm not convinced that does anything for a class A amp, but that's another story.
My first attack on hum is to move the output transformer away from the mains. That has helped in all five amps I've built or modded recently. In one case I even used a toroidal mains transformer, which certainly solves the stray field problem.
In one case I had to fit a C-L-C filter and pay out for a choke, and that also helps, but it adds cost, complication and weight to the amp. In other cases I have adopted the C-R-C filter as you say. This hum cancelling trick is a new one for me and I'm surprised how well it works. Clearly you need the right output transformer. My mini Princeton EL90 has such a transformer, and I intend to try the trick on that amp as well*
As to this "one valve" amp - my next session will concentrate on that FET front end. I want to see if I can tweak the bias to get a bit of crunch out of it without sacrificing gain. I have a plan, and that will be an evening with the signal generator and the scope I suspect. As always I design using test equipment, guesswork and ear. Finally I write down the circuit retrospectively. Huh. Engineers eh?
*If I buy an amp I am likely to leave it alone for quite some time. If however I have modded it or built it myself then the cover comes off for yet another experimental mod every few days. Maybe I should use velcro not screws 🙂
Pete
That purist attitude quickly goes out of the window when it comes to rectifiers. I use silicon and if I feel the need for sag, then the bridge diodes get 150 ohms in series. I'm not convinced that does anything for a class A amp, but that's another story.
My first attack on hum is to move the output transformer away from the mains. That has helped in all five amps I've built or modded recently. In one case I even used a toroidal mains transformer, which certainly solves the stray field problem.
In one case I had to fit a C-L-C filter and pay out for a choke, and that also helps, but it adds cost, complication and weight to the amp. In other cases I have adopted the C-R-C filter as you say. This hum cancelling trick is a new one for me and I'm surprised how well it works. Clearly you need the right output transformer. My mini Princeton EL90 has such a transformer, and I intend to try the trick on that amp as well*
As to this "one valve" amp - my next session will concentrate on that FET front end. I want to see if I can tweak the bias to get a bit of crunch out of it without sacrificing gain. I have a plan, and that will be an evening with the signal generator and the scope I suspect. As always I design using test equipment, guesswork and ear. Finally I write down the circuit retrospectively. Huh. Engineers eh?
*If I buy an amp I am likely to leave it alone for quite some time. If however I have modded it or built it myself then the cover comes off for yet another experimental mod every few days. Maybe I should use velcro not screws 🙂
Pete
I briefly owned a Fender Blues Junior. You could hear hum from the speaker with the mains transformer powered on - even when both output valves had been removed! Fender really blew it when they placed the power and output transformers so close together, and poorly oriented relative to each other on top of that.
And few of us will have those nowadays. So I'm curious: what happens if you tap a tiny bit of 60 Hz or 120 Hz AC (as appropriate for the hum frequency), from the power transformer secondary or the rectified output, and feed that into the input (or cathode of the input stage) with appropriate phasing? Can one do the same hum-cancelling trick that way, without needing the specialist transformer?
I know you have your own ideas to implement. But perhaps you might get some ideas out of this (post #43, JFET/MOSFET cascode, by MJD Tech): https://www.diyaudio.com/forums/instruments-and-amps/314730-tube-emulation-eq-5.html#post5263157
I know the feeling! Eventually I have to make myself stop by reminding myself that while fiddling with the amp, I've been neglecting to practice and play guitar, which was the whole point of building the guitar amp in the first place!
I'm going through a short break from electronics at the moment now, and spending time with the guitar instead. So instead of wasting my time on my electronics projects, I waste it sticking my big nose into your electronic projects... 😱
-Gnobuddy
Fender amps: I had to do some serious engineering on my Champ 600 and my Pro Junior. Looking inside I'm convinced Fender think a guitar amp should hum, otherwise it's not authentic. Grrr.
Hum cancelling: You would need to inject 100 (or 120)Hz from the rectifier. I have a feeling it would be tricky to get the exact phase and shape of signal though.
In fact the over wind on the primary is quite small. It could be possible to create your own by adding a new winding to the OPT with some enamel covered wire if there's room to thread it. It's a possibility...
Guitar: Yes, keep up the guitar practice 🙂 My method is to jam along to youtube backing tracks (Quist does good ones in all sorts of styles). His tracks repeat in the predictable manner but every now and then he throws in something odd and modal to make it interesting. My challenge is to think of something different to play over on each repeat.
On that subject. Is it permissible to put a sound sample of amplifiers up on the forum, as a MP3 attachment maybe. Does that work, it it the done thing?
Pete
Did you never part out old radios? I feel that at least at our side of the pond almost any post WWII radio had an output transformer with this hum compensation winding. Of course they aren't apt for HiFi, but surely will work well in a SE geetaar amplifier.
Best regards!
This hum cancellation method had been used for decades in table radios. Also it tends to cancel the DC in the trafo's core at the expense of lost signal AC in the series resistor.
Yes, I think it was a nice little trick to raise the quality of a cheap radio, and it works just fine with small guitar amps as well, if you can find the right transformer.
I have to say, my tiny transformer from a UK Murphy radio (U506 I think) has plenty of bass despite the small core.
At present the pentode clipping is a bit harsh for my tastes. I can't work out if that is the limits of the transformer or if it is the Pentode running out of peak current because I have it biassed a bit cold.
I have to say, my tiny transformer from a UK Murphy radio (U506 I think) has plenty of bass despite the small core.
At present the pentode clipping is a bit harsh for my tastes. I can't work out if that is the limits of the transformer or if it is the Pentode running out of peak current because I have it biassed a bit cold.
Or perhaps the clipping is because you are omitting the rest of the receiver current flowing though the compensating winding, so, now the transformer becomes saturated by this DC offset in the core.
I set the compensation winding current (using the associated resistor) by putting a pot in there and using the smallest resistance that also offered the lowest hum on the scope at the speaker.
The sharp clip is at the top (most positive) of the anode voltage trace, which I would have thought would be the minimum anode current. The bottom (max current) clip is more rounded. Perhaps the pentode/transformer are running out of voltage headroom?
To be honest then operation of the transformer under these extreme conditions is a bit of a mystery to me. It's a long time since I studied mutual inductance and similar theory.
The sharp clip is at the top (most positive) of the anode voltage trace, which I would have thought would be the minimum anode current. The bottom (max current) clip is more rounded. Perhaps the pentode/transformer are running out of voltage headroom?
To be honest then operation of the transformer under these extreme conditions is a bit of a mystery to me. It's a long time since I studied mutual inductance and similar theory.
The resistor is also loading the signal from the output pentode, so you are wasting sound(able) power in it, so, two things can happen: you need to increase signal until the pentode runs beyond the linear range, or the resistor loading the anode of the pentode shift the anode loading to an improper loadline.
This kind of trafo are designed to work in a well definite status, such as pentode load impedance, and DC flowing into the two opposite windings, so, changing any of those variables will invariably conduct to heavier distortion.
This kind of trafo are designed to work in a well definite status, such as pentode load impedance, and DC flowing into the two opposite windings, so, changing any of those variables will invariably conduct to heavier distortion.
Hi Osvaldo,
it's not the DC, it's the hum AC that is being (almost) cancelled by this method. or an effective DC cancellation the compensation winding's turns number is much too small, and so is the current flowing through it.
Best regards!
Partially true. What you are cancelling is [Amper * Turns], both DC and AC components are canceled or almost, severely attenuated.
When boy I disassembled some of them, and lacks the paper/air gap and generally putting all E's and all I's together creates a minimum gap necessary for proper operation.
When boy I disassembled some of them, and lacks the paper/air gap and generally putting all E's and all I's together creates a minimum gap necessary for proper operation.
Last edited:
Read this posting in the above mentioned UK Vintage Radio forum. I couldn't have said it better 😉.
The best hum cancellation is achieved if
Np/Nc = Rp/Rf
with Np is the plate winding turns number, Nc ist the cancellation winding turns number, Rp is the power tube's internal resistance, and Rf is the filter resistor's value. Best regards!
The best hum cancellation is achieved if
Np/Nc = Rp/Rf
with Np is the plate winding turns number, Nc ist the cancellation winding turns number, Rp is the power tube's internal resistance, and Rf is the filter resistor's value. Best regards!
Sure, you are forming a bridge. When balanced, no transference between opposite corners. But for DC, current flowing in opposite direction in both windings, is like a some kind of partial push pull with Ra in one side, and a R plus the DC flowing through the receiver in the other. I suspect that perfect cancelling (Bridge balance) can't be simultaneously in AC and DC components.
Yes, you're right. The only goal was hum cancellation, not DC balancing. Remember that these OT's still were gapped.
Best regards!
Best regards!
Yes, as already said, a natural gap between the two packs of laminations. I never found an added gap specifically, which doesn't imply that may be other with added gap material.
Thanks for the tip about Quist, I'll check him / her / it out!
I have a Digitech Trio+ I use to make (almost) instant backing tracks to play along with. It's a wonderful little tool.
When I want more creative control, I sometimes create my own backing tracks in Musescore (Free music composition and notation software | MuseScore )
I, for one, would love to hear some of your amp samples.
Quite ironically for an audio forum, MP3 is not one of the allowed file extensions for attachments. At least, it wasn't, the last time I checked.
However you can always rename your file - stick a .txt extension on the end or something like that, and that will allow you to upload it. Evidently the forum software only checks the filename extension, not the MIME type.
I've also seen people upload their track to one of the legal music sharing sites (Soundcloud, Bandcamp, Youtube, etc), and then provide a link here.
-Gnobuddy
With a pentode, you have at least three "knobs" to tweak: screen grid voltage, cathode resistor value, and the less obvious one: B+ voltage, which slides your transformer load-line horizontally left or right on top of the pentode curves.
So if you want a warmer bias but the power supply can't support any more quiescent bias current, you can lower the B+ voltage at the top end of the output transformer primary, which will slide the load line to the left, which will also lower the top left end of the load line, placing the existing bias point further up along it (i.e. you have a warmer bias point on your newly relocated load-line.)
I think you can do that with a resistor from power supply + to the + end of the OT primary to drop the extra voltage. You'll also need an electrolytic cap from that end of the primary to ground, sized to have much lower impedance than the OT primary itself at the lowest frequency of interest.
Of course you'll reduce maximum output power, but it may only be by a couple of decibels, which is insignificant. Getting the overdrive timbre you want is probably considerably more significant for your purposes as a musician.
-Gnobuddy
Just checked out Quist on You Tube. Geez, this guy is prolific!
Thank you, that was an excellent tip!
-Gnobuddy
I never have. Valve radios have been rare or non-existent in the places I've lived. When they do show up on Craigslist, they are not cheap, so I've never bought one.
In bigger cities, where I have been unfortunate enough to live most of my adult life, there is not much stability in people's lives. Few people have the chance to live and age gracefully in the same house for decades, surrounded by electronics they bought when they were young in the 1950s. Instead, most people have their lives constantly stirred up and move to new housing, new jobs, new relationships, new cities, every few years. Because of this, "vintage electronics" tends to mean something from the 1990s, not the 1950s. So I can find low quality cassette decks and "mini component systems" easily, but no valve radios!
Actually, I've only once ever stumbled across an affordable vintage piece of valve equipment. That was a Bogen gramaphone + P.A. system with a microphone input as well. Apparently it was originally designed for square dancing. The person running the dance put a record on the turntable, played the music through the P.A., and added his / her square dance calls using the microphone.
That Bogen is slumbering in our small basement storage area, waiting to be reborn as some sort of guitar amp one day.
-Gnobuddy
The hum-buck tap is less than 10% of the total. The buck-tap current is maybe 10% of the total. The DC cancellation can only be 1% at most.
I believe the hum-bucking has to be balanced against the values of the caps and the PSRR of the circuit, so just using one hum-buck OT in another circuit is not sure to get benefit.
I believe the hum-bucking has to be balanced against the values of the caps and the PSRR of the circuit, so just using one hum-buck OT in another circuit is not sure to get benefit.