I plan to build an amp with a single output impedance of 8 ohms. The designer has the winding spec's online. Is there anyway to just change the wire diameter and number of turns to get a 4 ohm output. That is much more in line with my speaker arrangement. At present there are 6 sections with 120 turns of .6 mm wire on a double bobbin 3 to the side.
Tad
Tad
Last edited:
tryonziess,
You will almost certainly have to derate ultimate sustained power a bit. You wont be able to make a direct trade as 8 ohms is not twice the turns of 4 ohms. Having said that, unless you are driving theatre sound systems your winder should be able to accomplish the change without audible problems. The thicker wire will provide a bigger build and this may further limit his choices. You should rely on his judgment, since he is the person tasked with actually winding the coils.
Bud
You will almost certainly have to derate ultimate sustained power a bit. You wont be able to make a direct trade as 8 ohms is not twice the turns of 4 ohms. Having said that, unless you are driving theatre sound systems your winder should be able to accomplish the change without audible problems. The thicker wire will provide a bigger build and this may further limit his choices. You should rely on his judgment, since he is the person tasked with actually winding the coils.
Bud
Tad,
With all respect, but you're talking basic transformer stuff here.
When there are 120 turns on the secondary (I guess 6 sections paralleled), and you know what the primary impedance is by checking the number of primary turns, it is quite simple to calculate the winding ratio for a 4 ohm load (not 8 ohms output impedance or 4 ohm output as you put it wrong).
OK, when there are 3000 primary turns, the primary impedance is 5k with an 8 ohm load (3000/120=25; 25 squared is 625; 625 x 8 = 5k.
For a 4 ohm load the number of secondary turns should be 85.
hey Tad,
if the 8ohm coils has 120turns, 4ohms will have 85 turns.....
going from 120 turns to 60 turns means going from 8ohms to 2ohms....
if you go thru the ARRL wire table i attached previously, you will see a column there with "turns per square inch", so that if you allow 80% of available window for your copper, you can estimate how many turns is possible with your copper wire size of choice...
YvesM had OPT designs with 6 ohms secondaries, i believe that such arrangement covers both 8 and 4 ohm speakers and i am doing that for my output traffo builds....
if the 8ohm coils has 120turns, 4ohms will have 85 turns.....
going from 120 turns to 60 turns means going from 8ohms to 2ohms....
if you go thru the ARRL wire table i attached previously, you will see a column there with "turns per square inch", so that if you allow 80% of available window for your copper, you can estimate how many turns is possible with your copper wire size of choice...
YvesM had OPT designs with 6 ohms secondaries, i believe that such arrangement covers both 8 and 4 ohm speakers and i am doing that for my output traffo builds....
Last edited:
Tad, you were going with what??
Do you plan to wind a copy of Chrisl Ludwig's output transformers?
When so, I'd say STOP RIGHT NOW and listen to me, because you're heading in the wrong direction (and magnet wire is expensive....). Before tossing around with numbers of turns a.s.o. you'd better learn some transformer basics first
.I checked Chrisl's website for his output transformer.
Plate-plate impedance is specified being 6k4/6k4.
Primary is 8 sections of 600 turns each (in series), total 4800 turns.
Secondary is 6 sections of 120 turns each (parallel).
Winding ratio primary to secondary is therefore 40:1 (4800/120).
Primary impedance is 40 squared x 8 = 12800 ohm (I prefer to specify plate to plate impedance being 12k8 instead of 6k4/6k4.
When copying the transformer for 4 ohms instead of 8 ohms, the number of turns for the secondary must be 85 (as stated before).
The point however is, Chrisl used a MD102b core (or SM102b) with a two chamber bobbin. Available winding width will be some 30mm per chamber; maximum winding heigth is 14mm for the used bobbin.
His wire gauge is more or less chosen for the available winding space .
For instance he winds each secondary section of 120 turns in three layers, and the primary sections in I guess 5 layers of 120 turns/layer.
This way, also taking care of winding space taken by layer insulation, he fills about all available winding space.
You must look at the available winding space with the cores / bobbins you plan to use.
Then you must compare iron area (Afe) of the MD102b core with what you have.
When necessary you must change the number of primary windings (and then of course also the number of secondary windings) to reach comparable primary induction and core excitation (depends on primary AC voltage).
Once you have done this it is very likely that you will need different wire gauges. Also you must calculate the winding space that your insulation foil takes.
So in the end it might well be that your coil differs from Chrisl's.
The double E core that Tony mentioned above uses special long E laminations (some outfit in NJ, had the patent I believe) with an (at least) double width back side to the E's. The opposing E's then get pushed all the way together without I lams. The back side will have somewhat reduced permeability compared to an I lam due to it's grain orientation. But the double E scheme does not suffer from a sudden drop in permeability when the flux reaches the 1/2 max point, where it would normally have to suddenly jump the E-I air gaps in an E-I setup. So the double E version is more constant permeability than E-I over the full flux range.
One could, I suppose, use I lams with the double E setup by using a normal stacking that leaves room for the I's. Then the benefits of both could be combined.
The main benefit is from using long E's, in either case, since they reduce leakage inductance considerably over scrapless E-I's. There are long E lams still commonly available (but without the doubled backside, see image attached) for constant voltage transformers.
One could, I suppose, use I lams with the double E setup by using a normal stacking that leaves room for the I's. Then the benefits of both could be combined.
The main benefit is from using long E's, in either case, since they reduce leakage inductance considerably over scrapless E-I's. There are long E lams still commonly available (but without the doubled backside, see image attached) for constant voltage transformers.
Attachments
Last edited:
Using the long aspect ratio Es is definitely a good way to help cut down on leakage. I use this dodge often with the right shape ferrite cores for SMPS transformers. Using a somewhat oversized stack to get the turns down also helps. Some people try using an undersized core and pile on the turns to compensate, and wonder why their transformer doesn't work all that well....
Thomas & Skinner has ferroresonant transformer lams, which have the long aspect ratio.
Thomas & Skinner has ferroresonant transformer lams, which have the long aspect ratio.
Hi there,
Here is a secondary arrangement which allows 4 or 8 or 16 (more exactly 4,9 and 16) with no compromise -understand: "using all the wire anyway".
Attachments
When choosing the C core for the outputs they rate them at 50 and 60 hertz by volt/amp capacity. How would this compare to the output transformer load. This amp uses 4 GU50 tubes for 160 watts output. Somewhat conservative and I wish the transformers to also be conservative in there capacity.
When looking at the bobbins on the Christl site and researching the MB102 transformers I was never able to get the exact deminsions of the wire window for the wind. The number of primary turns and the pri/sec ratio are all elements found from the TUBE data. The transformer core is selecting on its maximum core saturation. So if the total copper area does not completely fill the window does not thet allow for some extra room for output saturation. The larger bobbin also makes for for less layers.
I am aware this is much more complex than this explanation and that the first unit will most likely not be O-netics quality. It is just good that I have most of you to steer me away from a total catastrophe. I am also using C cores which are a little bit more efficient than the EI.
Is it not possible to use a properly layed out 250 watt output on a 160 watt amp?
And yes I have found that copper wire prices have gone up tremendously. Anyone ever have any luck salvaging copper wire from old motors?
Tad
When looking at the bobbins on the Christl site and researching the MB102 transformers I was never able to get the exact deminsions of the wire window for the wind. The number of primary turns and the pri/sec ratio are all elements found from the TUBE data. The transformer core is selecting on its maximum core saturation. So if the total copper area does not completely fill the window does not thet allow for some extra room for output saturation. The larger bobbin also makes for for less layers.
I am aware this is much more complex than this explanation and that the first unit will most likely not be O-netics quality. It is just good that I have most of you to steer me away from a total catastrophe. I am also using C cores which are a little bit more efficient than the EI.
Is it not possible to use a properly layed out 250 watt output on a 160 watt amp?
And yes I have found that copper wire prices have gone up tremendously. Anyone ever have any luck salvaging copper wire from old motors?
Tad
The power an OPT is able to carry for a given distortion depends on the frequency.
P=(F/K)² or K=F/sqr(P) or F=sqr(P)/K
Where P is the power in Watts, F is the frequency in Hz and K is specific for a given OPT.
From the published data, first find K: fe, if he claims for 10W at 30Hz, then K=30/sqr(10)=9.5.
Now, you can compute the admissible power at, say 20Hz: P=(20/9.5)²=4.4W.
Or the lowest frequency for 15W: F=sqr(15)/9.5=37Hz.
Differently said, the same OPT can be seen as:
- a 10W / 30Hz unit,
- a 4.4W / 20Hz unit,
- a 15W / 37Hz unit.
In your situation: K=50/sqr(250)=3.16
If you target is 30Hz, then you may expect P=(30/3.16)²)=90W . . . unless you abuse the iron !
Max power being limited anyway by the maximum voltages applied to any winding and by currents wich will flow into !
This does not apply to SE OPT where DC bias must also be taken into account.
Yves.
Hi Yves,
With this arrangement you effectively keep the B+ (which is ground in AC terms) and Loudspeaker Minus at close distance, which is good with regard to capacities.
The same for Anode connection and Loudspeaker Plus at the inside of the coil, closest to the core.
When I look at your secondary sections, they are each two layers, and each layer is bifilar wound.
So actually there are four 1 ohm sections within each of your three "two layer bifilar secondary" sections (four isolated windings of 28 turns).
This means that you could make the 4, 9 and 16 ohms arrangement also for each individual secondary section, and parallel these three sets of arrangements.
In this case B+ and anode connections are more spread over the coil with respect to the loudspeaker connections, which is different from your arrangement.
Did you try this?
When I was experimenting with these possible arrangements I found somewhat better HF frequency bandwidth with the "parallel" approach despite the series connection being better in theory because of less capactive coupling from ground to signal (but it's been quite a while ago so my memory is not 100% reliable
).
Hi YvesM,
when i found your OPT arrangement with 6 ohm only secondary, i took it to mean that is can be used for either 4 or 8 ohms,was i correct in assuming that?
tony
Hi Peter,
Indeed, the 4 by 1 arangement is by no mean a novelty, I've just "prewired" to reduce the number of terminals from 8 to 6 as well as the number of jumpers to move
Duh? Try ... what ? Please explain differently.
Yves.
Hi Peter,
Indeed, the 4 by 1 arangement is by no mean a novelty, I've just "prewired" to reduce the number of terminals from 8 to 6 as well as the number of jumpers to move
Duh? Try ... what ? Please explain differently.
Yves.
In your arrangemant you have just one Loudspeaker Minus terminal, which is close to B+, and also one Loudspeaker Plus terminal which is close to the anode terminal. That predicts good capacitive behaviour.
In the parallel arrangement (by making 4, 9 and 16 ohms within each secondary section there are three instead of one loudspeaker minus terminals at the coil, spread evenly over the coil, and also three loudspeaker plus terminals, also spread evenly over the coil.
That differs from your arrangement, of course also in capacitive behaviour.
Indeed 8 terminals are required for loudspeaker connection.
I was curious if you have tried this arrangement also.
http://www.dissident-audio.com/OPT_da/PST-and-OPT.tar.gz
Y'all.
I don't know how to open Yvesm's ta.gz file.
Anyone able to help?
thanks
tim
http://www.dissident-audio.com/OPT_da/PST-and-OPT.tar.gz
Y'all.
I don't know how to open Yvesm's ta.gz file.
Anyone able to help?
thanks
tim
http://www.dissident-audio.com/OPT_da/PST-and-OPT.tar.gz
Check and see if 7zip archive manager can handle tar files..(I believe it can) There should be several other archive programs that can unpack these files.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.