Going All the Way...
Okay everyone,
Here is a 2 stage version with phono line level input...
and a close up of the first stage drivers...
N.B. there is a LM317 lurking underneath the black heatsink.
Input transformer is a Gardners MU7566 audio line matching transformer - 600 ohm to 10 k ohm 1:4 step up. bandwidth - 1 dB about 20 Hz to 60 kHz acording to the data sheet.
The interstage transformer (which is the input transformer on the single stage version we have been discussing so far) is the 15VA toroid as previously described. The bi-filar wound 12volt windings allow it to work in this configuration.
There could be a long bit of wire between the driver and the 15VA transformer here which still allow the split between pre-amp and mono-block power amps next to the speakers.
Output is the transformer with the pretty colured mains cable windings. Hammering out bass quite nicely 🙂
Sounds okay although the distortion levels will (I assume) be a bit higher (may measure it some time).
This version of the amplifier could be driven directly from a DAC's current outputs.
BW,
Susan.
Okay everyone,
Here is a 2 stage version with phono line level input...
and a close up of the first stage drivers...
N.B. there is a LM317 lurking underneath the black heatsink.
Input transformer is a Gardners MU7566 audio line matching transformer - 600 ohm to 10 k ohm 1:4 step up. bandwidth - 1 dB about 20 Hz to 60 kHz acording to the data sheet.
The interstage transformer (which is the input transformer on the single stage version we have been discussing so far) is the 15VA toroid as previously described. The bi-filar wound 12volt windings allow it to work in this configuration.
There could be a long bit of wire between the driver and the 15VA transformer here which still allow the split between pre-amp and mono-block power amps next to the speakers.
Output is the transformer with the pretty colured mains cable windings. Hammering out bass quite nicely 🙂
Sounds okay although the distortion levels will (I assume) be a bit higher (may measure it some time).
This version of the amplifier could be driven directly from a DAC's current outputs.
BW,
Susan.
LOTS OF FUN WITH WINDINGS
I bought a E-150 transformer core and 3*10 meters of 1mm2 cable today. I miscalculated the amount of cable needed so I only got 43 turns trifilar winded. I plugged in the transformer to my circlotron amp to see what happend. It sounded great.Plenty of bass and treble too. I could not hear any nastyness att all. Tomorrow i will buy 20 meters more and fill the bobin as much as I can.
It seems to have some serious potential!!!!
Thanx Susan for all great idees and your sharing with us!!!
I bought a E-150 transformer core and 3*10 meters of 1mm2 cable today. I miscalculated the amount of cable needed so I only got 43 turns trifilar winded. I plugged in the transformer to my circlotron amp to see what happend. It sounded great.Plenty of bass and treble too. I could not hear any nastyness att all. Tomorrow i will buy 20 meters more and fill the bobin as much as I can.
It seems to have some serious potential!!!!
Thanx Susan for all great idees and your sharing with us!!!
Hi!
Hostmaster: thanks for the help, but are you sure?
I still didn't understand what bobo1on1 wrote:
"Connecting the speaker to the primaries means you will have to connect the speaker to the sources of the fets, because when you use a mains transformer you are using the secondaries as primaries."
Susan! You wrote:
"To maintain the coupling you may have to place the speaker across the primaries.
(You could also try using one as the input transformer if you don't have anything smaller.)", but in your last schem where is the coupling between the primary and secondary rails? and also where is the step down ratio 2:1 or 4:1 ? What is for the the 110 V rails (primary?) if we don't use? Is it just hang in the air?
Sorry for the stupid questions, but I can not understand this kind of transformator working. Until know I saw and understand only mains and Tube output transformers. :-((
DX.Master's revised amp wouldn't work as it is? I mean with the input and out put transformer arangement?
Thanks!
Tyimo
Hostmaster: thanks for the help, but are you sure?
I still didn't understand what bobo1on1 wrote:
"Connecting the speaker to the primaries means you will have to connect the speaker to the sources of the fets, because when you use a mains transformer you are using the secondaries as primaries."
Susan! You wrote:
"To maintain the coupling you may have to place the speaker across the primaries.
(You could also try using one as the input transformer if you don't have anything smaller.)", but in your last schem where is the coupling between the primary and secondary rails? and also where is the step down ratio 2:1 or 4:1 ? What is for the the 110 V rails (primary?) if we don't use? Is it just hang in the air?
Sorry for the stupid questions, but I can not understand this kind of transformator working. Until know I saw and understand only mains and Tube output transformers. :-((
DX.Master's revised amp wouldn't work as it is? I mean with the input and out put transformer arangement?
Thanks!
Tyimo
Tyimo said:
No, please see.
Post 379
Please see this picture, it's a redraw of the changes Susan made:
An externally hosted image should be here but it was not working when we last tested it.
I will construct the above Amp, i have finish 30% yet.
Best wishes - uwe
Re: LOTS OF FUN WITH WINDINGS
Hi Circlomanen,
Great to hear of your progress, many thanks for the update.
These bobbins do consume an inordinate amount of wire. I striped off 9-1/2 hand spans worth and even on the E-120 bobbin didn't have quite enough. Fortunately I had not cut them off the cable that I was sacrificing to the greater good of the DIYforum listees, so I was able to get another hand span's worth and complete the bobbin.
So I got about 60 turns per winding, and much better sound than I was expecting from such a quick hack. I made this version really to illustrate tri-filar winding and to see what sort of response one could get from doing this with rewinding mains transformers as this is been tried by bobo1on1 (and others) and I didn't feel I could make any categoric statements about the results without having done one myself.
Thank you. I think so too 🙂
I wouldn't listen to anything else although I would admit to the possibility of being biased having lived with the sound for ten years.
I do my best to try to explain "what and why" for my design work on this amplifier.
It is simple enough that it should be accessible to people with limited resources and who might not feel up to a wholesale 20 plus device (plus the 100 passives that go with) solid state amplifier or one of those valve thingies with all those scary high voltages (and I include myself in this statement as I designed this amplifier from first principles for that very reason).
And I believe that there is lots of room for experimentation with different transformer ratios, materials and mosfet types, etc. which is to me part and parcel what DIYaudio is about.
This is exemplified by the toroid version for example which I had mistakenly ignored after my own poor results of nearly ten years ago.
Look forward to hearing the results from you full bobbin.
Best wishes,
Susan.
Hi Circlomanen,
Circlomanen said:
Great to hear of your progress, many thanks for the update.
These bobbins do consume an inordinate amount of wire. I striped off 9-1/2 hand spans worth and even on the E-120 bobbin didn't have quite enough. Fortunately I had not cut them off the cable that I was sacrificing to the greater good of the DIYforum listees, so I was able to get another hand span's worth and complete the bobbin.
So I got about 60 turns per winding, and much better sound than I was expecting from such a quick hack. I made this version really to illustrate tri-filar winding and to see what sort of response one could get from doing this with rewinding mains transformers as this is been tried by bobo1on1 (and others) and I didn't feel I could make any categoric statements about the results without having done one myself.
Thank you. I think so too 🙂
I wouldn't listen to anything else although I would admit to the possibility of being biased having lived with the sound for ten years.
I do my best to try to explain "what and why" for my design work on this amplifier.
It is simple enough that it should be accessible to people with limited resources and who might not feel up to a wholesale 20 plus device (plus the 100 passives that go with) solid state amplifier or one of those valve thingies with all those scary high voltages (and I include myself in this statement as I designed this amplifier from first principles for that very reason).
And I believe that there is lots of room for experimentation with different transformer ratios, materials and mosfet types, etc. which is to me part and parcel what DIYaudio is about.
This is exemplified by the toroid version for example which I had mistakenly ignored after my own poor results of nearly ten years ago.
Look forward to hearing the results from you full bobbin.
Best wishes,
Susan.
Hi Tyimo,
The schematic as now drawn - see:
http://bauteile-fuer-die-elektronik.de/projekte/Zeus/images/constant-current--2-1024.png
... with the low voltage windings being used with the loudspeaker across them is correct for when using a mains toroid transformer as the output inductor.
This is not my prefered arrangment as it does not allow optimal load coupling - however it is more than adequate to get going with and to have something running on the bench.
The step down ratio is for my multi-filar wound output transformer (see my website for details) which provides an effective 2:1 or selectable 2:1/4:1 reduction depending on the version.
This better matches the impedances but at the cost of raw power - which is not really what this amplifier is about but it is still gratifying that some people (amp-man) have happily pushed 250 watts.
Hope this helps.
Best wishes,
Susan.
Tyimo said:
The schematic as now drawn - see:
http://bauteile-fuer-die-elektronik.de/projekte/Zeus/images/constant-current--2-1024.png
... with the low voltage windings being used with the loudspeaker across them is correct for when using a mains toroid transformer as the output inductor.
This is not my prefered arrangment as it does not allow optimal load coupling - however it is more than adequate to get going with and to have something running on the bench.
The step down ratio is for my multi-filar wound output transformer (see my website for details) which provides an effective 2:1 or selectable 2:1/4:1 reduction depending on the version.
This better matches the impedances but at the cost of raw power - which is not really what this amplifier is about but it is still gratifying that some people (amp-man) have happily pushed 250 watts.
Hope this helps.
Best wishes,
Susan.
Hi,
Thanks for your post and the update to your schematic - all is hopefully now clear.
I was busy composing my replies off line as I use the spelling checker in my email program to try to sort out those "interesting" words. A blessing for those of us who are dyslexic.
Look forward to hearing your impressions once you have it running.
Are you building a pair, or just the one?
Best wishes,
Susan.
Thanks for your post and the update to your schematic - all is hopefully now clear.
I was busy composing my replies off line as I use the spelling checker in my email program to try to sort out those "interesting" words. A blessing for those of us who are dyslexic.
dx.master said:
Look forward to hearing your impressions once you have it running.
Are you building a pair, or just the one?
Best wishes,
Susan.
Hi Susan,
I hope too.
I start with just one 'simple' Amp with a simple wood housing case.
But I have the parts for the second too.
See:
http://bauteile-fuer-die-elektronik.de/projekte/Zeus/
Susan-Parker said:
I hope too.
I start with just one 'simple' Amp with a simple wood housing case.
But I have the parts for the second too.
See:
http://bauteile-fuer-die-elektronik.de/projekte/Zeus/
Hi Uwe,
Nice and practical.
To my chigran I just have the current test version of my dual stage amplifier strung out in bits across my desk in front of the flat screen monitor - but I am running off a bench PSU (30 volt - 500 mA ) so there are no exposed mains voltages floating around.
Even with "inadequate" heatsinks the mosfet thermal tracking on the gate bias seems to be keeping everything stable.
Thanks for the pictures - it's good to see.
Best wishes,
Susan.
dx.master said:
Nice and practical.
To my chigran I just have the current test version of my dual stage amplifier strung out in bits across my desk in front of the flat screen monitor - but I am running off a bench PSU (30 volt - 500 mA ) so there are no exposed mains voltages floating around.
Even with "inadequate" heatsinks the mosfet thermal tracking on the gate bias seems to be keeping everything stable.
Thanks for the pictures - it's good to see.
Best wishes,
Susan.
People in North America who may want Hitachi TO3 case mosfets, matched, I have some. Quite honestly, there are less expensive devices to be had. But these are curve traced matched and have the metal TO3 cases, which are thermally perhaps better than the plastic pak devices. 2SJ49
Also have a few toroidal power xfmrs in the 2 x 20 vac secondaries, for mains use, not output use (imho these are not bifilar wound and don't have the requisite freq response) @~275VA. They will produce ~35vdc when filtered. Obviously you can run the two secondaries in parallel for this amp.
Please contact me off list via email: bear@bearlabs.com
Hope this is ok to post... <duck>
_-_-bear
Also have a few toroidal power xfmrs in the 2 x 20 vac secondaries, for mains use, not output use (imho these are not bifilar wound and don't have the requisite freq response) @~275VA. They will produce ~35vdc when filtered. Obviously you can run the two secondaries in parallel for this amp.
Please contact me off list via email: bear@bearlabs.com
Hope this is ok to post... <duck>
_-_-bear
Hi Susan!
Thanks a lot!
Wouldn't be possible to multi-filar wound a torroid for output transformer? Like Plitron do?
Tyimo
Thanks a lot!
Wouldn't be possible to multi-filar wound a torroid for output transformer? Like Plitron do?
Tyimo
Susan-Parker said:
Yes, I think the push-pull makes a great difference. Although the input Z of the Hiraga is 10K or so, ALL the hum (even less now with a toroid mains xformer) comes from inductive coupling to the output transpormer. Actually, I confirmed that disconnecting the amp but just let the power xformer running still caused the hum... Eerie, a "zero-transistor amp" that humms. Probably because it doen't know the words (I know, an old one. Couln't resist😀 )
Jan Didden
Hi Tyimo,
Ta.
I see no reason why one couldn't strip off the windings from a toroid transformer and then tri-filar wind one's own on the core.
Don't forget to keep some of the insulation as the core edges can be quite sharp.
It would be a bit fiddly using enamelled single core heavy copper wire but using mains cable conductors as I have done recently (or even similar wire purchased on seperate reels) it should be easy enough if a bit tedious.
It may be a bit fatter than before, but I am quite confident that if you can get 60 turns or so for the trifilar windings on say a 200 to 350 VA size core you should get something that works well enough to start with.
Just like making a bobble to go on a knitted hat, except one doesn't do the cutting bit at the end to make it all fluffy 🙂
Best wishes,
Susan.
Tyimo said:
Ta.
I see no reason why one couldn't strip off the windings from a toroid transformer and then tri-filar wind one's own on the core.
Don't forget to keep some of the insulation as the core edges can be quite sharp.
It would be a bit fiddly using enamelled single core heavy copper wire but using mains cable conductors as I have done recently (or even similar wire purchased on seperate reels) it should be easy enough if a bit tedious.
It may be a bit fatter than before, but I am quite confident that if you can get 60 turns or so for the trifilar windings on say a 200 to 350 VA size core you should get something that works well enough to start with.
Just like making a bobble to go on a knitted hat, except one doesn't do the cutting bit at the end to make it all fluffy 🙂
Best wishes,
Susan.
Hi Jan,
Thanks for that 🙂
Even if it is an old one it did make me laugh.
Best wishes,
Susan.
janneman said:
Thanks for that 🙂
Even if it is an old one it did make me laugh.
Best wishes,
Susan.
Hallo Uwe!
Danke für Dich auch! Jetzt ist alles klar!
grüsse vom Tyimo
Hi Susan!
Thanks again!!
Greets:
Tyimo
Danke für Dich auch! Jetzt ist alles klar!
grüsse vom Tyimo
Hi Susan!
Thanks again!!
Greets:
Tyimo
A QUESTION...
If I want a amp with good bassrespons (say -3 db at 10 Hz 100w output) i should have many turns of winding (and a BIG transformercore) and if I want good treble its fewer turns but thicker copperwire????
What differens does it make if I use a singel thicker or many thinner wires parallel in each winding for a given primary or secondary and the same amount of turns???
(oops thats two questions!)
If I want a amp with good bassrespons (say -3 db at 10 Hz 100w output) i should have many turns of winding (and a BIG transformercore) and if I want good treble its fewer turns but thicker copperwire????
What differens does it make if I use a singel thicker or many thinner wires parallel in each winding for a given primary or secondary and the same amount of turns???
(oops thats two questions!)
Tyimo said:
Bitte kein Problem.
Der Test-Amp ist nun fertig.
The Test-Amp is now ready for testing.
Please look at:
http://bauteile-fuer-die-elektronik.de/projekte/Zeus/
- uwe
To Susan:
Susan, I would like to email you something of interest. Would you be so kind to mail me (through the forum) an email address I could use?
Thanks,
Jan Didden
Susan, I would like to email you something of interest. Would you be so kind to mail me (through the forum) an email address I could use?
Thanks,
Jan Didden
Re: A QUESTION...
Hi Circlomanen,
Thank you for your questions 🙂
Okay:
========
Question 1:
Because we are using saturatable magnetic cores there is a trade off between size and power.
I look at a core and by rule of thumb have worked out that for a wideband audio transformer one should have 4 times the VA rating as the equivalent line 50/60 Hz transformer for a given power.
E.g.
For a 50 watt transformer use a 200 VA core.
For a 100 watt transformer us a 400 VA core.
However this is for a -3 dB at 25 Hz response.
To go to 12.5 Hz at the same power one would have to DOUBLE the VA rating, so a 100 watt output would require an 800 VA core.
Also the power required starts to increase at the very bottom end as the DC and near DC resistance of the windings come into play rather than the AC inductance.
There are two separate things to consider here in the transformer core.
1. The core area gives the maximum flux density that the transformer can handle before saturating.
Note: Mains power transformer's are often optimized to run at up to a 80% to 85% or higher flux at their maximum rating (which is why a 50 Hz transformer can work at 60 Hz but a 60 Hz optimized transformer sometimes can't be run at 50 Hz because).
For audio use I want to run the transformers well below their saturation point so that one is not getting into the curls at the top and bottom of the flux curves.
Remember flux density is produced by current - not voltage.
2. The number of turns through the length of the closed part of the magnetic circuit gives the inductance.
I.e. Using the same size laminations in both cases; a lamination stack of 2 inches in depth with 100 turns has a similar inductance to a lamination stack of 4 inches with 50 turns.
The 2 inch lamination stack may have about 50% of the winding length in the magnetic circuit, the 4 inch stack could have about 75%. The 4" stack would have double the flux capacity.
With regard for the difference between low end optimized and high end optimized one generally needs far less power for the higher frequencies (we are talking personal HiFi here rather than sound reinforcement for public outdoor events) and thus I would not see the need for thicker wire.
You can use fewer turns but I would caution about over-optimizing as one still needs to keep the transformer stable at lower frequencies even if they are at reduced levels.
A lot depends on your crossover point. Mine are in the 100 to 200 Hz range, but if you were driving a horn treble and a big 12" bass/midrange then these things would be different.
-
So if you wanted to use a toroid core then a fat dumpy one is better than a thin wide one, all other things being equal.
To increase power handling you could take two thin wide cores and stack them together and then wind around both.
========
Question 2:
Using one thick or several thin wires together is the same as far as inductance is concerned. However you are reducing the cross section and therefor the current capacity of the wire by using several instead of one for a given overall space.
BUT, this may not be so bad as the tri-filar mains flex conductor wound transformer test shows there is quite a lot of leeway in how one goes about winding these transformers.
For high power there comes a point where one would want to use multiple mosfets. The best way of connecting them to avoid current hogging is for each to be connected to a separate thinner winding rather than paralleled across one really thick one.
For example I am looking at trying out the use of two 0.71 mm wires instead of one 1.0 mm wire, so the transformer primaries would have four wires in total, two per side, each with it's own mosfet.
The secondaries could also be several thiner wires but bear in mind that here one wants as low a DC resistance as possible to stop that uncontrolled ported bass driver from flapping (did I mention anywhere that I am not so keen on ported loudspeakers?).
The problem here of course is that there comes a point where trying to multi-filar wind large numbers of different sized wires becomes problematic, especially when one is doing this by hand.
It is very important that all the primaries are multi-filar wound to get good close coupling. By using multiple interleaved layers one can of course couple the primaries and the secondaries - but at the expense of simplicity and ease of construction - and some performance loss. It then becomes more of a conventional valve style device, with a correspondingly higher price tag if one is having them made.
=======
I hope this answers your question(s)?
Best wishes,
Susan.
Hi Circlomanen,
Thank you for your questions 🙂
Circlomanen said:
Okay:
========
Question 1:
Because we are using saturatable magnetic cores there is a trade off between size and power.
I look at a core and by rule of thumb have worked out that for a wideband audio transformer one should have 4 times the VA rating as the equivalent line 50/60 Hz transformer for a given power.
E.g.
For a 50 watt transformer use a 200 VA core.
For a 100 watt transformer us a 400 VA core.
However this is for a -3 dB at 25 Hz response.
To go to 12.5 Hz at the same power one would have to DOUBLE the VA rating, so a 100 watt output would require an 800 VA core.
Also the power required starts to increase at the very bottom end as the DC and near DC resistance of the windings come into play rather than the AC inductance.
There are two separate things to consider here in the transformer core.
1. The core area gives the maximum flux density that the transformer can handle before saturating.
Note: Mains power transformer's are often optimized to run at up to a 80% to 85% or higher flux at their maximum rating (which is why a 50 Hz transformer can work at 60 Hz but a 60 Hz optimized transformer sometimes can't be run at 50 Hz because).
For audio use I want to run the transformers well below their saturation point so that one is not getting into the curls at the top and bottom of the flux curves.
Remember flux density is produced by current - not voltage.
2. The number of turns through the length of the closed part of the magnetic circuit gives the inductance.
I.e. Using the same size laminations in both cases; a lamination stack of 2 inches in depth with 100 turns has a similar inductance to a lamination stack of 4 inches with 50 turns.
The 2 inch lamination stack may have about 50% of the winding length in the magnetic circuit, the 4 inch stack could have about 75%. The 4" stack would have double the flux capacity.
With regard for the difference between low end optimized and high end optimized one generally needs far less power for the higher frequencies (we are talking personal HiFi here rather than sound reinforcement for public outdoor events) and thus I would not see the need for thicker wire.
You can use fewer turns but I would caution about over-optimizing as one still needs to keep the transformer stable at lower frequencies even if they are at reduced levels.
A lot depends on your crossover point. Mine are in the 100 to 200 Hz range, but if you were driving a horn treble and a big 12" bass/midrange then these things would be different.
-
So if you wanted to use a toroid core then a fat dumpy one is better than a thin wide one, all other things being equal.
To increase power handling you could take two thin wide cores and stack them together and then wind around both.
========
Question 2:
Using one thick or several thin wires together is the same as far as inductance is concerned. However you are reducing the cross section and therefor the current capacity of the wire by using several instead of one for a given overall space.
BUT, this may not be so bad as the tri-filar mains flex conductor wound transformer test shows there is quite a lot of leeway in how one goes about winding these transformers.
For high power there comes a point where one would want to use multiple mosfets. The best way of connecting them to avoid current hogging is for each to be connected to a separate thinner winding rather than paralleled across one really thick one.
For example I am looking at trying out the use of two 0.71 mm wires instead of one 1.0 mm wire, so the transformer primaries would have four wires in total, two per side, each with it's own mosfet.
The secondaries could also be several thiner wires but bear in mind that here one wants as low a DC resistance as possible to stop that uncontrolled ported bass driver from flapping (did I mention anywhere that I am not so keen on ported loudspeakers?).
The problem here of course is that there comes a point where trying to multi-filar wind large numbers of different sized wires becomes problematic, especially when one is doing this by hand.
It is very important that all the primaries are multi-filar wound to get good close coupling. By using multiple interleaved layers one can of course couple the primaries and the secondaries - but at the expense of simplicity and ease of construction - and some performance loss. It then becomes more of a conventional valve style device, with a correspondingly higher price tag if one is having them made.
=======
I hope this answers your question(s)?
Best wishes,
Susan.