Hi guys, here's one for you valve/magnetics wizards. If it has been covered to death I apologise and hang my head but I couldn't find anything really close to my query.
I have a peculiar requirement which I am sure is going to generate "why on Earth do it that way" responses from some. It's a real problem I'm trying to solve but it's also an interesting area for me as I have little expertise in transformer design. My electronics knowledge is excellent, I've been a design engineer in the audio and seismic fields all my (long) life so the principles are there but the practical experience and knowledge of ironmongery is lacking.
I run a H&K TM36 guitar amp which has an output of 8-16ohm without any requirement to match. I know this is unusual but it's the way guitar amps seem to be going today. My speakers are 2x8ohm with the thought that I could run them in series at 16ohm and match. It works fine.
That said, I have always had an overly toppy and slightly brittle feel so I ran it with one speaker at 8ohm. It sounded much smoother, so I guessed that it responded to the damping factor improvement and the amp liked that level of load a little better than up at the top. Now I'm not thrilled at the prospect of having to buy two new speakers at close to £300 to rematch at 8ohm. So I became intrigued at the prospect of getting a matching transformer to do the job for me, but I can't find that sort of thing anywhere. Next thought was - design and make my own. Ok, I admit it, the worm became buried deep in the brain at this point!
My thoughts were twofold, (love that sort of old word). Firstly, as the input is coming from the output Tx of the amp it is already fully isolated, an autotransformer would be a good choice. 4ohm - 8ohm; 1:2; 1:1.414 winding ratio; 1.5:1 would be as good giving me a little bit of design leeway a nominal 10ohm at the amp output. An autotransformer should be easy to wind to that spec. But how many turns for the input and on what core? I believe auto-Txs are much easier going on the core than standard primary/secondary types so a suitable toroidal core should be ok to find. Can anyone cast any light on whether this is a sensible approach and where to look for materials?
Secondly I wondered, does a standard power transformer exist which would do the job already? Considering 2x115V primaries in parallel I would need about 75V/80V at the secondary. A 2x40V tranny is possible, the two primaries in parallel and the two secondaries in series give me a 1.4375:1 ratio and 8.25ohm (on paper). The lowest frequency I will be using is about 82Hz so a 50Hz at the requisite VA would seem ok for core size and lower limit. It's the upper frequency range which is another unknown to me. I know that a power Tx core is not ideally suited to this work but just how bad does it get?
I still lean towards the AutoTx idea as I know this lessens the requirements on the core a lot but the fine detail is beyond me at present, and I have to say that the online tutorials on this shed no real light on the solutions, they just regurgitate their versions of the formulae without any insight. Formulae I can do but I don't like applying anything without a feel for what it really implies in the real world.
Has anyone played with this idea or can you offer any guidance?
I have a peculiar requirement which I am sure is going to generate "why on Earth do it that way" responses from some. It's a real problem I'm trying to solve but it's also an interesting area for me as I have little expertise in transformer design. My electronics knowledge is excellent, I've been a design engineer in the audio and seismic fields all my (long) life so the principles are there but the practical experience and knowledge of ironmongery is lacking.
I run a H&K TM36 guitar amp which has an output of 8-16ohm without any requirement to match. I know this is unusual but it's the way guitar amps seem to be going today. My speakers are 2x8ohm with the thought that I could run them in series at 16ohm and match. It works fine.
That said, I have always had an overly toppy and slightly brittle feel so I ran it with one speaker at 8ohm. It sounded much smoother, so I guessed that it responded to the damping factor improvement and the amp liked that level of load a little better than up at the top. Now I'm not thrilled at the prospect of having to buy two new speakers at close to £300 to rematch at 8ohm. So I became intrigued at the prospect of getting a matching transformer to do the job for me, but I can't find that sort of thing anywhere. Next thought was - design and make my own. Ok, I admit it, the worm became buried deep in the brain at this point!
My thoughts were twofold, (love that sort of old word). Firstly, as the input is coming from the output Tx of the amp it is already fully isolated, an autotransformer would be a good choice. 4ohm - 8ohm; 1:2; 1:1.414 winding ratio; 1.5:1 would be as good giving me a little bit of design leeway a nominal 10ohm at the amp output. An autotransformer should be easy to wind to that spec. But how many turns for the input and on what core? I believe auto-Txs are much easier going on the core than standard primary/secondary types so a suitable toroidal core should be ok to find. Can anyone cast any light on whether this is a sensible approach and where to look for materials?
Secondly I wondered, does a standard power transformer exist which would do the job already? Considering 2x115V primaries in parallel I would need about 75V/80V at the secondary. A 2x40V tranny is possible, the two primaries in parallel and the two secondaries in series give me a 1.4375:1 ratio and 8.25ohm (on paper). The lowest frequency I will be using is about 82Hz so a 50Hz at the requisite VA would seem ok for core size and lower limit. It's the upper frequency range which is another unknown to me. I know that a power Tx core is not ideally suited to this work but just how bad does it get?
I still lean towards the AutoTx idea as I know this lessens the requirements on the core a lot but the fine detail is beyond me at present, and I have to say that the online tutorials on this shed no real light on the solutions, they just regurgitate their versions of the formulae without any insight. Formulae I can do but I don't like applying anything without a feel for what it really implies in the real world.
Has anyone played with this idea or can you offer any guidance?
Hi,
The damping factor is worse driving 8 ohms rather than 16 ohms.
The sensible would just add another pair of 12s for an 8 ohm 4x12.
That will effectively increase max SPL by 6dB, that is apparently
quadrupling the amplifier output if it likes 8 ohm loading.
rgds, sreten.
The damping factor is worse driving 8 ohms rather than 16 ohms.
The sensible would just add another pair of 12s for an 8 ohm 4x12.
That will effectively increase max SPL by 6dB, that is apparently
quadrupling the amplifier output if it likes 8 ohm loading.
rgds, sreten.
Last edited:
I think what H&K have done is make the output stages and Tx so over engineered they are effectively bomb proof. They specifically state that the output will drive 8-16ohm, and that's without taps. Now that has to represent a mismatch of course, but the electronics seems to be able to cope with it. It's just that I have heard both 16ohm and 8ohm now and it sounds a little more balanced with a 8ohm load than with 16ohm.
As we know the damping factor is the speaker's coil resistance (not it's total impedance) divided by the external resistance across it. The lower that external resistance is, the higher the damping factor. While the damping factor will improve for a single 16ohm driver over a single 8ohm, it doesn't work like that in my case. It's definitely not better at 16ohm with two speakers in series when viewed from each individual speaker's perspective.
With just a single 8ohm speaker the damping factor depends on only: amp output Tx
With two 8ohm speakers in series the impedance each speaker sees its signal coming from is: amp output Tx + the other speaker. The DF has to have dropped below 1.
With the two speakers in parallel they each see an impedance of: amp output Tx // the other speaker. This is lower and leads to a higher DF again.
It is that last setup, 2 x 8ohm in parallel, which I would like to use and hence was curious about the matching autotransformer approach. I agree that I could replace the speakers or add 2 more for a 4x12" approach, but both are expensive and the 4 x 12" is a damned sight more awkward to carry around. My current two 1 x 12" cabs are the most convenient setup I have ever had.
As we know the damping factor is the speaker's coil resistance (not it's total impedance) divided by the external resistance across it. The lower that external resistance is, the higher the damping factor. While the damping factor will improve for a single 16ohm driver over a single 8ohm, it doesn't work like that in my case. It's definitely not better at 16ohm with two speakers in series when viewed from each individual speaker's perspective.
With just a single 8ohm speaker the damping factor depends on only: amp output Tx
With two 8ohm speakers in series the impedance each speaker sees its signal coming from is: amp output Tx + the other speaker. The DF has to have dropped below 1.
With the two speakers in parallel they each see an impedance of: amp output Tx // the other speaker. This is lower and leads to a higher DF again.
It is that last setup, 2 x 8ohm in parallel, which I would like to use and hence was curious about the matching autotransformer approach. I agree that I could replace the speakers or add 2 more for a 4x12" approach, but both are expensive and the 4 x 12" is a damned sight more awkward to carry around. My current two 1 x 12" cabs are the most convenient setup I have ever had.
Hi,
The DF does not work in the way you describe, each speaker
does not damp the other, DF for 16 ohm is double that for 8 ohm.
Similarly parallel drivers don't improve matters, its half that for 8 ohm.
Going used you should be able to swap the 8 ohm drivers
to 16 ohm for very little net cost, which is what you should
have got in the first place, two 16 ohm cabinets, for parallel.
Maximum output power for two cabs, reduced a bit for one.
rgds, sreten.
The DF does not work in the way you describe, each speaker
does not damp the other, DF for 16 ohm is double that for 8 ohm.
Similarly parallel drivers don't improve matters, its half that for 8 ohm.
Going used you should be able to swap the 8 ohm drivers
to 16 ohm for very little net cost, which is what you should
have got in the first place, two 16 ohm cabinets, for parallel.
Maximum output power for two cabs, reduced a bit for one.
rgds, sreten.
two 16ohm cabinets coupled in parallel, or one cabinet with two 16ohms drivers in parallel, or one cabinet with two 4ohms drivers wired in series.
Or two 16ohms cabinets each with two 8ohms driver wired in series, then wired in parallel for an effective 8ohms loading with twice the Sd and roughly twice the SPL.
Another 16ohms cabinet with dual 8ohms drivers would be ideal.
Or two 16ohms cabinets each with two 8ohms driver wired in series, then wired in parallel for an effective 8ohms loading with twice the Sd and roughly twice the SPL.
Another 16ohms cabinet with dual 8ohms drivers would be ideal.
What you are looking for is an auto-transformer with a 1:1.4 turns ratio, much cheaper I suspect than getting a bunch of additional speakers despite the potential advantages of such an approach if you actually need the additional efficiency/spls.
BTW this really belongs in Instruments & Amps since the question involves a guitar amp. I will move it.
BTW this really belongs in Instruments & Amps since the question involves a guitar amp. I will move it.Hi,
Except an audio 1:1.4 transformer basically does not exist. There is
no demand or need for them in any practical commercial product.
Multiple output taps do the job if its needed for valve amplifiers.
rgds, sreten.
Except an audio 1:1.4 transformer basically does not exist. There is
no demand or need for them in any practical commercial product.
Multiple output taps do the job if its needed for valve amplifiers.
rgds, sreten.
So if I have an amp with an output impedance of say 1ohm and I drive a 6ohm (resistance) speaker with it, the damping factor is 6, (6ohm because the coil resistance is less than the quoted 8ohm impedance). What would the effect on the damping factor be to add a series 5ohm resistor between amp and speaker, (like an extreme bad speaker cable)?
So on to speakers rather than resistors. I have attached a bit of advice I found on the subject from JBL. They advise against series connection just for this very reason. If you skip to the ringed part you can see they say, "...connecting separate woofers in series reduces the damping factor of the amplifier to a value less than 1." How does that happen then?
http://www.diyaudio.com/forums/attachment.php?attachmentid=488506&stc=1&d=1434301447
A 1:1.4 autotransformer can be bought off the shelf and relatively cheaply if you look at 25V, 70V and 100V line transformers and do a bit of calculation. Or you can always wind your own, (if you know how to choose a core and decide on the number of turns on the primary). So I'm back to my original post. (I think Kevink got where I was with that. 😉 )
I take the point about selling my own on and replacing with 16ohm types, it's just that the worm is in my head now and I want to understand this approach before I bin it. I'm sure you will understand that weakness.
Thanks for moving the post Kevin, sorry for misplacing it in the first place.
So on to speakers rather than resistors. I have attached a bit of advice I found on the subject from JBL. They advise against series connection just for this very reason. If you skip to the ringed part you can see they say, "...connecting separate woofers in series reduces the damping factor of the amplifier to a value less than 1." How does that happen then?
http://www.diyaudio.com/forums/attachment.php?attachmentid=488506&stc=1&d=1434301447
A 1:1.4 autotransformer can be bought off the shelf and relatively cheaply if you look at 25V, 70V and 100V line transformers and do a bit of calculation. Or you can always wind your own, (if you know how to choose a core and decide on the number of turns on the primary). So I'm back to my original post. (I think Kevink got where I was with that. 😉 )
I take the point about selling my own on and replacing with 16ohm types, it's just that the worm is in my head now and I want to understand this approach before I bin it. I'm sure you will understand that weakness.
Thanks for moving the post Kevin, sorry for misplacing it in the first place.
Attachments
Hi,
The JBL datasheet is simply very wrong, end of story.
(Its laughably inept, given common real experience.)
Adding 5R to your case will reduce DF from 6 to 1.
Why is it wrong ? For the same reason you can connect
dual coiled subs in series as well as parallel, there is
no electrical difference between dual coils and separate
dual drivers, in series and parallel.
I'm well aware you can get 1:1.4 transformers for line
systems, but try finding them for 8 ohm to 16 ohm,
at the sort of power you presumably need.
You will need a transformer the same size as the
amplifiers output transformer, and the same quality.
Swapping your drivers is the painless usual route.
Or flog one 1x12 and the others driver and put a
monster 8 ohm in one cabinet, like an Electrovoice.
Or learn to live with one 8 ohm cabinet.
rgds, sreten.
The JBL datasheet is simply very wrong, end of story.
(Its laughably inept, given common real experience.)
Adding 5R to your case will reduce DF from 6 to 1.
Why is it wrong ? For the same reason you can connect
dual coiled subs in series as well as parallel, there is
no electrical difference between dual coils and separate
dual drivers, in series and parallel.
I'm well aware you can get 1:1.4 transformers for line
systems, but try finding them for 8 ohm to 16 ohm,
at the sort of power you presumably need.
You will need a transformer the same size as the
amplifiers output transformer, and the same quality.
Swapping your drivers is the painless usual route.
Or flog one 1x12 and the others driver and put a
monster 8 ohm in one cabinet, like an Electrovoice.
Or learn to live with one 8 ohm cabinet.
rgds, sreten.
Last edited:
Hmmm! I have to say when I read your reply I was still a bit sceptical sreten. I spent another hour or so online looking at many references to this and the concerted opinion was that you're wrong and the second speaker resistance DOES make a difference. Then I found it, the Holy Grail, a clear statement of the true situation with proper analysis and experimental results! And waddya know, you're absolutely 100% right!!!!
I've attached the pdf of the article which made it all clear. I appreciate your time in posting a couple of times in the face of my lack of comprehension, (and your patience with my questioning). I'm sure you'll appreciate my desire to understand the whys rather than just blindly accepting the hows. I did mention this is also a learning experience for me.
However.... (There had to be a however didn't there?)
If I find a line transformer with a 1.4:1 turns ratio, (initially I did say 1:1.4 but I'm sure you get the point here), and the output tap is 4ohms, (my 2 x 8ohm in parallel), how is that not a 8ohm to 4ohm matching setup? (And you've now established that theoretically it would also work just as well the other way around with the speakers in series for 16ohm output impedance.)
There are transformers out there like this. For example, I found the AtlasSound AF140 which claims voltage ratio 1.4, Impedance Ratio 1.96 and is rated at 150W. There is the Electrovoice AT100 which diagramatically shows a tap at 4ohm and a tap at 8ohm with nominal turns ratio referenced to 4ohm winding of 1.41. Are these types of autotransformer not suitable for this work? Frequencywise we are only putting a guitar signal through it, 80Hz - 5kHz, and I would have thought this should be on a par with 70V line type applications.
Or even my initial light hearted suggestion of a 2x115V/2x40V power transformer? 115V in parallel and 40V in series. 115:80 ~ 1.44 turns ratio, 2.01 impedance ratio.
I've attached the pdf of the article which made it all clear. I appreciate your time in posting a couple of times in the face of my lack of comprehension, (and your patience with my questioning). I'm sure you'll appreciate my desire to understand the whys rather than just blindly accepting the hows. I did mention this is also a learning experience for me.
However.... (There had to be a however didn't there?)
If I find a line transformer with a 1.4:1 turns ratio, (initially I did say 1:1.4 but I'm sure you get the point here), and the output tap is 4ohms, (my 2 x 8ohm in parallel), how is that not a 8ohm to 4ohm matching setup? (And you've now established that theoretically it would also work just as well the other way around with the speakers in series for 16ohm output impedance.)
There are transformers out there like this. For example, I found the AtlasSound AF140 which claims voltage ratio 1.4, Impedance Ratio 1.96 and is rated at 150W. There is the Electrovoice AT100 which diagramatically shows a tap at 4ohm and a tap at 8ohm with nominal turns ratio referenced to 4ohm winding of 1.41. Are these types of autotransformer not suitable for this work? Frequencywise we are only putting a guitar signal through it, 80Hz - 5kHz, and I would have thought this should be on a par with 70V line type applications.
Or even my initial light hearted suggestion of a 2x115V/2x40V power transformer? 115V in parallel and 40V in series. 115:80 ~ 1.44 turns ratio, 2.01 impedance ratio.
Over what frequency range?
Iron cored transformer lose a lot of their power capability as the frequency goes down.
100Hz performance could be excellent @ 150W into 4ohms.
try 50Hz, is it still excellent, or good, or just about OK
try 30Hz, is is OK, or mediocre, or bad?
Your rewinding idea is better.
Introducing a 4ohm tap at ~ 70% of the output winding will reduce your power capability slightly, but you're still using your existing transformer and using all of the existing primary.
The AF140 claims 30Hz-15kHz +-1dB Andrew. They're not exactly cheap units but not overly expensive. They're a damned sight cheaper than two new drivers, (but it may of course come to that option if I can't solve the parts and process quandry). For guitar work the bottom note is 82Hz, I'm not a bass player nor do I have the hereditary sixth finger for a seventh string, and it doesn't really extend much beyond 5kHz. Of course there is always some energy outside of this range but it is limited.
I should have said earlier, I don't have an AF140 or an A100 in my possession, they're just potential models I came up with in my search. I'm still trying to get advice which will let me see the best way to solve this problem. Unfortunately these things all seem to be in 'Americee' and priced very reasonably until you factor in the cost of postage for iron core units!!!!
I also found a really promising range of GOSS toroidal O-cores which would have been "spiffing".
Toroidal O-Core OA-80
When I tested the water by putting that one in my basket I found that it cost just under $15 and it would have cost $110 on top of that to post here!
One thing I'm making sure of is to go for units which are rated at around 100W even though the amp I'm trying to interface with only chucks out 36W. I don't know too much Tx fine detail but I gather one problem is, (and it makes sense to me), that when the core saturates and no further magnetic flux change can be generated it loses the ability to resist increased current and effectively shorts out the input. I gather it occurs at the lower limits of the unit's frequency range. I don't like the sound of that at all!
I should have said earlier, I don't have an AF140 or an A100 in my possession, they're just potential models I came up with in my search. I'm still trying to get advice which will let me see the best way to solve this problem. Unfortunately these things all seem to be in 'Americee' and priced very reasonably until you factor in the cost of postage for iron core units!!!!
I also found a really promising range of GOSS toroidal O-cores which would have been "spiffing".
Toroidal O-Core OA-80
When I tested the water by putting that one in my basket I found that it cost just under $15 and it would have cost $110 on top of that to post here!
One thing I'm making sure of is to go for units which are rated at around 100W even though the amp I'm trying to interface with only chucks out 36W. I don't know too much Tx fine detail but I gather one problem is, (and it makes sense to me), that when the core saturates and no further magnetic flux change can be generated it loses the ability to resist increased current and effectively shorts out the input. I gather it occurs at the lower limits of the unit's frequency range. I don't like the sound of that at all!
The frequency response +-1dB will probably be for ~1W
The 150W frequency response will be very different. And I don't know nothing about valve/tube amplifiers !
Then buy one that has 16ohms tapping.
That is the easiest solution.
There is no point in buying the wrong amplifier !
The 150W frequency response will be very different. And I don't know nothing about valve/tube amplifiers !
Amplifier not bought yet.
Then buy one that has 16ohms tapping.
That is the easiest solution.
There is no point in buying the wrong amplifier !
Last edited:
No, I meant that I don't have the transformer yet. The amp is a Hughes and Kettner Tubemeister 36 which is an amazing sounding amp for guitar. This has a single output into 8-16ohm without the need for taps or accurate matching. It's just that, having tried it on both, I think it sounds a little better with 8ohm on it that 16ohm.
Take your point about the frequency range specified at 1W. That would make it look better wouldn't it?
Take your point about the frequency range specified at 1W. That would make it look better wouldn't it?
Weber offers an impedance converter transformer, WZC-100. It is an auto transformer with taps for 2, 2.6, 4, 5.3, 8 and 16 Ohms.
Find it here: Weber Speakers - Making the world a bit louder each day.
Find it here: Weber Speakers - Making the world a bit louder each day.
That looks like it might be an auto-transformer.
Since the Valve amp output is already isolated from the HV side, then an auto transformer would be OK as the impedance converter.
Makes me wonder if it might be much easier to wind your own auto-transformer using a core from any big recovered EI. (the 300VA to 400VA core from a 900W microwave would be a good start).
Since the Valve amp output is already isolated from the HV side, then an auto transformer would be OK as the impedance converter.
Makes me wonder if it might be much easier to wind your own auto-transformer using a core from any big recovered EI. (the 300VA to 400VA core from a 900W microwave would be a good start).
The Weber unit is *designed* to do exactly that and nothing else, go figure.
And yes, you can wind your own, problem is designing the transformer from the ground up, and then having the necessary hardware to actually wind it.
That said, and sorry for this, I believe you get better sound (at least to your ears) by using a single speaker instead of two, but much doubt it's damping related, so I guess you are barking at the wrong tree.
If anything, I believe with a single speaker you are driving your amp harder and getting closer to its "sweet spot".
And yes, you can wind your own, problem is designing the transformer from the ground up, and then having the necessary hardware to actually wind it.
That said, and sorry for this, I believe you get better sound (at least to your ears) by using a single speaker instead of two, but much doubt it's damping related, so I guess you are barking at the wrong tree.
If anything, I believe with a single speaker you are driving your amp harder and getting closer to its "sweet spot".
- Status
- Not open for further replies.