I've always assumed that a smoothing choke with a fairly high inductance would be a pretty large beast.
By checking out places like Mouser Electronics etc. this is obviously not so.
A smoothing choke from Triad Magnetics of 10Henries @ 5omA and Rdc=500 ohms is only 1.625" wide, 2.812" tall and 1.5" deep.
https://no.mouser.com/datasheet/2/410/C_3X-1892608.pdf
This seems almost too good to be true . . . or is this a badly designed choke?
By checking out places like Mouser Electronics etc. this is obviously not so.
A smoothing choke from Triad Magnetics of 10Henries @ 5omA and Rdc=500 ohms is only 1.625" wide, 2.812" tall and 1.5" deep.
https://no.mouser.com/datasheet/2/410/C_3X-1892608.pdf
This seems almost too good to be true . . . or is this a badly designed choke?
This choke uses thin wire, so you get a lot of turns (more turns = more inductance) in small space. The price one pays for it is that the RDC is high. OTOH, power supplies using chokes with too little RDC and small caps can cause unstable supplies, so that high RDC is not necessarily bad. Blencowe explains that well in his book Designing Power Supplies for Valve amplier, 2nd edition.
Chokes from eg Monolith magnetics are also not much larger https://www.monolithmagnetics.com/sites/default/files/ps chokes datasheet.pdf
Chokes from eg Monolith magnetics are also not much larger https://www.monolithmagnetics.com/sites/default/files/ps chokes datasheet.pdf
Tim de Paravicini: As a rule of thumb the choke must be twice as large (physically) as the power transformer.
-A beast in other words.
I've got a huge number of pdf-files on choke and OT-winding, but this video is the best I've come across so far.
This is somewhat correct in one specific case at least.
The pic below shows my heavily modded Dynaco PAS3 preamp. From right to left, the silvery thingy labelled "PA 522" is an aftermarket power trafo after the original one burned. The 2 yellowish ones are chokes made in Germany, their sizes are similar to the one the OP posted. I initially tried a choke input with one of those, and it buzzed, very loudly. Tried silent blocs, they didn't help. I finally bought the left most one, from China, and it's alright. The configuration is LC + 2 x (LC).
The moral of the story is: a LC topology is BRUTAL on the head choke: it has to comply with electrical rules such as critical inductance, and also mechanical ones, such as robust construction, and oversizing.
The pic below shows my heavily modded Dynaco PAS3 preamp. From right to left, the silvery thingy labelled "PA 522" is an aftermarket power trafo after the original one burned. The 2 yellowish ones are chokes made in Germany, their sizes are similar to the one the OP posted. I initially tried a choke input with one of those, and it buzzed, very loudly. Tried silent blocs, they didn't help. I finally bought the left most one, from China, and it's alright. The configuration is LC + 2 x (LC).
The moral of the story is: a LC topology is BRUTAL on the head choke: it has to comply with electrical rules such as critical inductance, and also mechanical ones, such as robust construction, and oversizing.
Last edited:
I am listening to it now, a lot of good info on OPTs, still waiting for the info on the chokes. Where is that info?
I saw that part now. He is talking about choke input supplies, there indeed the choke must be big and probably none of the Triad stuff will work there. If you go for a capacitor input supply (about 95% of amps use this), the triad will do its job.
Simply said,
The more turns a choke has for a specific core, the more flux density headroom you have to spend on the 3 following parameters.
-Choke inductance
-Maximum Idc
-Maximum Vrms at a minimum frequency.
The more you need one of them, the more you'll have to sacrifice the rest of them.
For example, if a choke is specified at 10H @ 100mA, the same can be also gapped at 5H @ 200mA, resulting in the same flux density.
For choke input duty, you'll have to consider a high amount of Ripple Vrms across the choke, where the frequency is the double of the mains frequency (100 and 120Hz).
Using a choke can be flexible. If you intend using one for choke input duty while exceeding the quoted saturation Vrms, you can back down on Idc for compensation. You can't go back on inductance, as the choke is probably already gapped and assembled. However, if the choke can be disassembled, you could alter the inductance by changing the airgap.
However, you need to know the choke's innards, such as the turn count, the core Area and its magnetic properties, at least an idea of saturation flux density, which also varies by the airgap change. As a rule of thumb, I find that GOSS begins "kneeling" above 1.4T, Hitachi amorphous C-cores at 1T and nanocrystalline C cores at 0.85T.
Higher turn count however, in order to fit within the window, has to employ thinner wire, hence the higher Rdc. The main Rdc drawback is the heat losses of the choke and the higher voltage drop, so an Rdc to Idc balance is mandatory and this varies from one manufacturer to the other.
A bigger choke potentially means the following:
-Higher core surface area -> Higher flux density headroom vs lower Rdc
-Higher windows area -> Higher flux density headroom vs lower Rdc
The more turns a choke has for a specific core, the more flux density headroom you have to spend on the 3 following parameters.
-Choke inductance
-Maximum Idc
-Maximum Vrms at a minimum frequency.
The more you need one of them, the more you'll have to sacrifice the rest of them.
For example, if a choke is specified at 10H @ 100mA, the same can be also gapped at 5H @ 200mA, resulting in the same flux density.
For choke input duty, you'll have to consider a high amount of Ripple Vrms across the choke, where the frequency is the double of the mains frequency (100 and 120Hz).
Using a choke can be flexible. If you intend using one for choke input duty while exceeding the quoted saturation Vrms, you can back down on Idc for compensation. You can't go back on inductance, as the choke is probably already gapped and assembled. However, if the choke can be disassembled, you could alter the inductance by changing the airgap.
However, you need to know the choke's innards, such as the turn count, the core Area and its magnetic properties, at least an idea of saturation flux density, which also varies by the airgap change. As a rule of thumb, I find that GOSS begins "kneeling" above 1.4T, Hitachi amorphous C-cores at 1T and nanocrystalline C cores at 0.85T.
Higher turn count however, in order to fit within the window, has to employ thinner wire, hence the higher Rdc. The main Rdc drawback is the heat losses of the choke and the higher voltage drop, so an Rdc to Idc balance is mandatory and this varies from one manufacturer to the other.
A bigger choke potentially means the following:
-Higher core surface area -> Higher flux density headroom vs lower Rdc
-Higher windows area -> Higher flux density headroom vs lower Rdc
Last edited:
It would seem, that if I pull off my amp building plans further than just dreaming about it, I will be the happy owner of the heaviest 2 watt flea amp in the northern hemisphere.
In the car world "there's no replacement for displacement" . . . same is true for transformers and chokes;- there's no replacement for weight . . .
In the car world "there's no replacement for displacement" . . . same is true for transformers and chokes;- there's no replacement for weight . . .
Tell us about your plans! 🙂
Just heard Tim not being positive about C-Cores, still I am going to present you to this choke by Lundahl that weighs 1.35kg, can work in choke input (no personal experience, but I saw multiple people using LL in choke input supplies), has enough current capability for 2 Watt amps (thinking about something as 2x 2A3) and enough inductance to provide clean DC.
https://www.audiohobby.eu/en/lundah...-lundahl-power-supply-choke-ll2742-100ma.html
Just heard Tim not being positive about C-Cores, still I am going to present you to this choke by Lundahl that weighs 1.35kg, can work in choke input (no personal experience, but I saw multiple people using LL in choke input supplies), has enough current capability for 2 Watt amps (thinking about something as 2x 2A3) and enough inductance to provide clean DC.
https://www.audiohobby.eu/en/lundah...-lundahl-power-supply-choke-ll2742-100ma.html
Last edited:
My plans?
Monoblock flea amp (two off for stereo) #45 output, #26 input.
Regulated DC heater voltages for the #26's. (Rod Coleman's filament regs.)
Hardwired exept for the DC heater voltage supply.
Lundahl mic. transformers on the inputs of the #26's to help with the gain.
Choke input powersupplies.
Completely separate powersupplies for the #45 and #26 which means the #26 input stages get their own mains transformers,
tube rectifiers, chokes and PSU filters . . . the whole works.
I plan to wind all the iron myself (except the mic transformers of course), otherwise I could never afford all that iron.
I have successfully rewound the mains and output transformers for a VOX AC30 guitar amp and also a few other mains transformers
so I'm not completely unfamiliar with the challenges of transformer winding . . . and there will be challenges I fear.
I may be wrong, but my philosphy (at least in my own head) is that an amplifier no matter how good parts you use won't sound any better
than the quality of the PSU.
Those amps will possibly become insanely heavy for the one and a half watts they give in return.
I have a few Coral Flat 8ii full range driver units (100dB) so low wattage doesn't worry me much.
So, these days I'm reading up on all the PDF-books on transformer winding I can get my hands on.
This is going to take a long time . . .
Monoblock flea amp (two off for stereo) #45 output, #26 input.
Regulated DC heater voltages for the #26's. (Rod Coleman's filament regs.)
Hardwired exept for the DC heater voltage supply.
Lundahl mic. transformers on the inputs of the #26's to help with the gain.
Choke input powersupplies.
Completely separate powersupplies for the #45 and #26 which means the #26 input stages get their own mains transformers,
tube rectifiers, chokes and PSU filters . . . the whole works.
I plan to wind all the iron myself (except the mic transformers of course), otherwise I could never afford all that iron.
I have successfully rewound the mains and output transformers for a VOX AC30 guitar amp and also a few other mains transformers
so I'm not completely unfamiliar with the challenges of transformer winding . . . and there will be challenges I fear.
I may be wrong, but my philosphy (at least in my own head) is that an amplifier no matter how good parts you use won't sound any better
than the quality of the PSU.
Those amps will possibly become insanely heavy for the one and a half watts they give in return.
I have a few Coral Flat 8ii full range driver units (100dB) so low wattage doesn't worry me much.
So, these days I'm reading up on all the PDF-books on transformer winding I can get my hands on.
This is going to take a long time . . .
That is quite a plan, and I agree that it will take time. Until now, I kept away from winding, except some low voltage windings on existing toroidal transformers, for filaments or bias supplies. I never did a choke input supply as I do not own appropriate chokes, and they are rather expensive to buy, so I stayed with Cap input supplies. From my reading I understand that they have better regulation, which is good for class AB or class B amplifiers (less sag), so I also do not see much advantage in using them in my amps, which are mostly class A or very much in class A. But the day I see some nice Lundahl chokes at the swap meet I may bite.
So, going back to the original question. If your plans are for choke input, the Triad choke you found is indeed too good to be true for this particular application.
So, going back to the original question. If your plans are for choke input, the Triad choke you found is indeed too good to be true for this particular application.
A class A amp has almost no variation in current so a cap input supply will work, but consider this: A choke input supply is much
kinder to the mains transformer and the rectifier since the rectifier will now conduct through the full 180 degrees of the cycle,-
provided the choke is properly dimensioned. Also the ripple voltage will no longer be a sawtooth but a nice sinuswave which will eliminate
high frequency noise.
A cap input will produce a sawtooth ripple and nasty current spikes in the rectifier, and if the input cap is large enough the current
spikes could exceed the cathode rating for the tube rectifier.
All in all, a choke input supply will be more silent and give a longer life to the mains transformer and the rectifier.
It seems that tube rectifiers were made to work with either a choke input supply or a cap input supply with a very small input cap.
When we hear of people who have problems with tube rectifiers dying too soon or arcing the cause is probably a cap input supply.
As for the Triad choke, it seems to be far too small. I suspect that a good choke will be quite large and heavy.
kinder to the mains transformer and the rectifier since the rectifier will now conduct through the full 180 degrees of the cycle,-
provided the choke is properly dimensioned. Also the ripple voltage will no longer be a sawtooth but a nice sinuswave which will eliminate
high frequency noise.
A cap input will produce a sawtooth ripple and nasty current spikes in the rectifier, and if the input cap is large enough the current
spikes could exceed the cathode rating for the tube rectifier.
All in all, a choke input supply will be more silent and give a longer life to the mains transformer and the rectifier.
It seems that tube rectifiers were made to work with either a choke input supply or a cap input supply with a very small input cap.
When we hear of people who have problems with tube rectifiers dying too soon or arcing the cause is probably a cap input supply.
As for the Triad choke, it seems to be far too small. I suspect that a good choke will be quite large and heavy.
A cheap quick way to get a choke is to find a big EI mains tfmr, strip it down to the bobbin and laminations, figure out what gauge wire you want to handle the current, then just wind as many turns as you can fit on the bobbin. Re-assemble with all the "I"'s on top, pop in some thickish card for a gap and you have a choke. No calculations, just loads of turns. Measure the inductance and bish bash bosh jobs a good un.
I know this a bit crude but it works if your on a budget and large EI mains tfmr's can be had cheap.
Andy.
I know this a bit crude but it works if your on a budget and large EI mains tfmr's can be had cheap.
Andy.
I know all the arguments in favor of the choke input PS. One can say that the burden on PT and rectifiers is now passed to the choke, which will also produce spikes that can induce noise in other magnetic parts. And it is a special choke, more expensive, big, more exotic. As said, I will probably try it if I get a 2nd hand choke, but it must not absolutely be so for a class A amplifier. Also, in the heydays of chokes (and swinging chokes) capacitors were very expensive, so filtering was done with chokes, this can be done differently (and cheaper) now.
The objective here is fun, however, and maybe getting things done. A story: about 15 years ago I bought a set of 2nd hand Sowter OPTs with 10k/120mA, and some GM70s. I looked at these for all these years, moved them a couple of times, thinking: I must put these to use. I do not doubt that a choke input PS with mercury rectifiers would be great for it, but I did not have those parts and did not want to collect even more stuff in my man's cave. In the end I decided that for me KISS was best to get this thing running, and that is what I did. I made the PS with about 10 solder points, the concept is described here https://www.diyaudio.com/community/threads/something-interesting-class-a-hybrid.403286/#post-7453206 The amps are now with a friend who likes them better than a EAR 859 from Tim. And I am happy that I finally did it, and can now think about improvements.
The triad choke you found is just that, call it a black box that will give you filtering corresponding to 10H (-20,+50%) when pulling up to 50mA through it, with an internal resistance of 500R. It will buzz and scream loudly if you want to pull 50mA through it in a choke input PS, but it was not made for that.
A PS with low DCR chokes and small value caps can also misbehave quite badly, read Merlin Blencowe about it. Also, this text by DHTRob has some interesting information https://www.dhtrob.com/overige/pdf/dhtrob_psu.pdf
Good luck!
Erik
The objective here is fun, however, and maybe getting things done. A story: about 15 years ago I bought a set of 2nd hand Sowter OPTs with 10k/120mA, and some GM70s. I looked at these for all these years, moved them a couple of times, thinking: I must put these to use. I do not doubt that a choke input PS with mercury rectifiers would be great for it, but I did not have those parts and did not want to collect even more stuff in my man's cave. In the end I decided that for me KISS was best to get this thing running, and that is what I did. I made the PS with about 10 solder points, the concept is described here https://www.diyaudio.com/community/threads/something-interesting-class-a-hybrid.403286/#post-7453206 The amps are now with a friend who likes them better than a EAR 859 from Tim. And I am happy that I finally did it, and can now think about improvements.
The triad choke you found is just that, call it a black box that will give you filtering corresponding to 10H (-20,+50%) when pulling up to 50mA through it, with an internal resistance of 500R. It will buzz and scream loudly if you want to pull 50mA through it in a choke input PS, but it was not made for that.
A PS with low DCR chokes and small value caps can also misbehave quite badly, read Merlin Blencowe about it. Also, this text by DHTRob has some interesting information https://www.dhtrob.com/overige/pdf/dhtrob_psu.pdf
Good luck!
Erik
The ripple through the choke in choke input duty will lead to higher mechanical movement, hence noise, so a better building quality is required for such chokes.
If I had to buy quality iron for these plans I could just as well give it all up. It would have become too expensive.
Another thought I had was concerning the term diminishing returns.
Audio seems to be the perfect stage for throwing money, time and in this case iron in an endless roundabout of diminishing returns.
But, as I'm planning to wind the iron myself it won't be too expensive, and I need something to do in the wintertime up here in the arctic anyway.
As for the chokes I do think they need to be vacuum impregnated in order to make them mechanically solid.
I remember my first try on winding an output transformer many years ago.
Little did I understand that 240 volts AC through a coil with an iron core inside equals mechanical force.
After winding the bobbin I inserted the laminations, didn't bother with bolting them down and gave it 240 volts of juice from the wall socket.
The resulting buzz was enormously loud and the laminations flew out of the bobbin like rockets.
After bolting it down it was still buzzing like a beehive, so mechanical solidness is a must.
Another thought I had was concerning the term diminishing returns.
Audio seems to be the perfect stage for throwing money, time and in this case iron in an endless roundabout of diminishing returns.
But, as I'm planning to wind the iron myself it won't be too expensive, and I need something to do in the wintertime up here in the arctic anyway.
As for the chokes I do think they need to be vacuum impregnated in order to make them mechanically solid.
I remember my first try on winding an output transformer many years ago.
Little did I understand that 240 volts AC through a coil with an iron core inside equals mechanical force.
After winding the bobbin I inserted the laminations, didn't bother with bolting them down and gave it 240 volts of juice from the wall socket.
The resulting buzz was enormously loud and the laminations flew out of the bobbin like rockets.
After bolting it down it was still buzzing like a beehive, so mechanical solidness is a must.