Ampeg V4
- Thread starter CJ
- Start date
Help Support GroupDIY Audio Forum:
we want to remove sec. turns to drop B+,
we have 540 plate, we need around 480 plate volts DC,
we do not want to unwind the sec to count turns, easier just to remove a few from the outside layer, usually the pri is wound first on a pwr trans, so we do not have to take apart the whole coil to get to the sec,
but we still need to know the total turns of the sec to figure out how many turns to take off the winding to drop the AC voltage
we use the PRR method of determining turns without unwinding,
this involves squeaking 9 or 10 turns past the coil and core so that we create another winding on the transformer, if there is enuff room to do this, we can plug the xfmr into 120 primary and read the volts on the aux wind,
then we know 3 of 4 variable in the turns equation which means we can solve for sec turns,
first we determine the RMS voltage (AC)that yields the stock 540 DC volts, so we measure the AC volts that go into the bridge rectifier circuit, we find that value to be 399 V-ac, why not just divide our desired DC plate voltage of 480 by 1.414, the root of 2, (peak volts DC = V-ac x 1.414) to find the new V-ac that we want on the sec after removing turns?
we could, but there are losses associated with transformers, DCR of the winds, core loss, etc, so depending on the transformer and how big the core and wire is, this ratio of AC to DC will be different than 1.414,
for this beast, it turns out to be 540/399=1.353, pretty close to 1.414 as the wire is big and has low DCR as not many turns are required on the big 175 EI core,
so now we determine what AC voltage we want from the winding in order to get our desired B+ voltage of 480, so we use the ratio constant we determined above,
540 DC needs 399 AC, so 480 DC needs 480/1.353 = 355 V-ac sec volts,
so how may turns do we take off the sec to get 355?
well, with 119 V-ac applied to the pri, we get 5.94 V-ac from our test wind of 9 turns,
now we can figure out the pri turns, 119/5.94 = 20.4
so call the turns ratio 20:1
so to get pri turns we multiply our test wind by 20,
9T x 20 = 180 pri turns,
since we know pri turns and voltage along with sec voltage, we can find sec turns,
399 sec V-ac/119 Pri V-ac = 3.353, so now we are almost there,
since our pri to sec turns ratio 3.353 ,
3.353 x 180 = 603.54 sec turns so call it 604 turns,
so we have 180 T pri and 604T sec, we want to reduce this 604 turns by the ratio of the AC volts that we want dropped,
so 604 turns is to 399 V-ac as X turns is to 355 V-ac, our desired sec voltage which will give 480 DC for the plates the 6L6 tubes,
so reduce the 604 turns by a ratio of 355/399=0.89,
so 0,89 x 604 = 537 turns
so we take off 604-537= 67 Turns
we could have done it like this also: 480 B+/540 B+ = 89 percent,
.89 x 604=537, 604-537=67 turns ,
here is the coil with the test wind on it>
we have 540 plate, we need around 480 plate volts DC,
we do not want to unwind the sec to count turns, easier just to remove a few from the outside layer, usually the pri is wound first on a pwr trans, so we do not have to take apart the whole coil to get to the sec,
but we still need to know the total turns of the sec to figure out how many turns to take off the winding to drop the AC voltage
we use the PRR method of determining turns without unwinding,
this involves squeaking 9 or 10 turns past the coil and core so that we create another winding on the transformer, if there is enuff room to do this, we can plug the xfmr into 120 primary and read the volts on the aux wind,
then we know 3 of 4 variable in the turns equation which means we can solve for sec turns,
first we determine the RMS voltage (AC)that yields the stock 540 DC volts, so we measure the AC volts that go into the bridge rectifier circuit, we find that value to be 399 V-ac, why not just divide our desired DC plate voltage of 480 by 1.414, the root of 2, (peak volts DC = V-ac x 1.414) to find the new V-ac that we want on the sec after removing turns?
we could, but there are losses associated with transformers, DCR of the winds, core loss, etc, so depending on the transformer and how big the core and wire is, this ratio of AC to DC will be different than 1.414,
for this beast, it turns out to be 540/399=1.353, pretty close to 1.414 as the wire is big and has low DCR as not many turns are required on the big 175 EI core,
so now we determine what AC voltage we want from the winding in order to get our desired B+ voltage of 480, so we use the ratio constant we determined above,
540 DC needs 399 AC, so 480 DC needs 480/1.353 = 355 V-ac sec volts,
so how may turns do we take off the sec to get 355?
well, with 119 V-ac applied to the pri, we get 5.94 V-ac from our test wind of 9 turns,
now we can figure out the pri turns, 119/5.94 = 20.4
so call the turns ratio 20:1
so to get pri turns we multiply our test wind by 20,
9T x 20 = 180 pri turns,
since we know pri turns and voltage along with sec voltage, we can find sec turns,
399 sec V-ac/119 Pri V-ac = 3.353, so now we are almost there,
since our pri to sec turns ratio 3.353 ,
3.353 x 180 = 603.54 sec turns so call it 604 turns,
so we have 180 T pri and 604T sec, we want to reduce this 604 turns by the ratio of the AC volts that we want dropped,
so 604 turns is to 399 V-ac as X turns is to 355 V-ac, our desired sec voltage which will give 480 DC for the plates the 6L6 tubes,
so reduce the 604 turns by a ratio of 355/399=0.89,
so 0,89 x 604 = 537 turns
so we take off 604-537= 67 Turns
we could have done it like this also: 480 B+/540 B+ = 89 percent,
.89 x 604=537, 604-537=67 turns ,
here is the coil with the test wind on it>
Attachments
no, not really, we are just dropping the volts, wire size is the same, so current delivery will be the same, actually a little better with the reduced DCR from less turns, there is a DCR of only 9 ohms on the stock sec wind,
it turns out that they wound the 6.3 heater winding last, so we need to take it off to get at the sec so we can strip our 67 turns,
we find that the heater wind has 10 turns, we can check the accuracy of the PRR method as there are sometimes linkage issues involved depending on the transformer being tested,
we measured 6.58 V-ac on our heater winding, unloaded,
since we get 5.94 volts on our 9 turn test winding, what should be our V-ac for 10 turns?
5.94/9=.66 volts per turn,
so 10 turns should yield .66 x 10 = 6.6 volts, we measured 6.58, pretty dang close,
this gives us some confidence while removing the 67 turns,
it turns out that they wound the 6.3 heater winding last, so we need to take it off to get at the sec so we can strip our 67 turns,
we find that the heater wind has 10 turns, we can check the accuracy of the PRR method as there are sometimes linkage issues involved depending on the transformer being tested,
we measured 6.58 V-ac on our heater winding, unloaded,
since we get 5.94 volts on our 9 turn test winding, what should be our V-ac for 10 turns?
5.94/9=.66 volts per turn,
so 10 turns should yield .66 x 10 = 6.6 volts, we measured 6.58, pretty dang close,
this gives us some confidence while removing the 67 turns,
Attachments
Nice CJ, I don't know if you have made a recipe for doing the transformer rework. Would be nice to have in case we need to do so, I've seen a lot of yours here, but a step by step with cares and tips would be awesome, maybe is already there but I haven't seen it.
It could, in theory, deliver more current, could be a problem for the wire size. In this case I wouldn't bother much, the change is about 10% so inside the error, plus, in a HV tube circuit this still isn't a big deal, no much current compared to other factors. I would be careful in a lab PS with high current capability, able to drive short circuits, which takes abuse and must survive. The higher current may affect the peak current in first caps in the line, if they are too big, which may cause a problem, since lower resistivity, but as I said, won't be a problem in this case. If you are taking half the sec you probably will be better changing the sec for a higher gauge and rewinding it from scratch, for a 10% it doesn't worth it.
JS
Dreams said:
It could, in theory, deliver more current, could be a problem for the wire size. In this case I wouldn't bother much, the change is about 10% so inside the error, plus, in a HV tube circuit this still isn't a big deal, no much current compared to other factors. I would be careful in a lab PS with high current capability, able to drive short circuits, which takes abuse and must survive. The higher current may affect the peak current in first caps in the line, if they are too big, which may cause a problem, since lower resistivity, but as I said, won't be a problem in this case. If you are taking half the sec you probably will be better changing the sec for a higher gauge and rewinding it from scratch, for a 10% it doesn't worth it.
JS
under the 10 turn heater wind we find the green sec wind,
wire size looks like about a #23, our AWG chart says 1/2 amp can come out of this wire and this chart is very conservative, so figure 1, probably 2 amps with no heating at all,
we removed the turns and the V-ac is right around 355 which should make the tubes last longer and sound better, and we have no bulky power resistor heating up the chassis and wasting power, we did the bake and varnish dip to prevent any further rusting and now are wiring this amp back up,
wire size looks like about a #23, our AWG chart says 1/2 amp can come out of this wire and this chart is very conservative, so figure 1, probably 2 amps with no heating at all,
we removed the turns and the V-ac is right around 355 which should make the tubes last longer and sound better, and we have no bulky power resistor heating up the chassis and wasting power, we did the bake and varnish dip to prevent any further rusting and now are wiring this amp back up,
Attachments
one aspect if removing turns is B-max, flux will go up when turns are reduced, but since B+ is also coming down to 480, flux will be coming down also as the peak AC voltage will be reduced, and this is a monster core anyway, compared to most pwr transformers,
this guy runs at about 12 kG at 60 CPS,
a 2 inch stack of 175 EI was used, 110 lams = 0.018" instead of the standard .014"
Lap 1 instead of the usual Lap 3,
low DCR means a stiff power source, no sag, less compressed sound, power amps modulate the pwr supply, so pwr transformers can influence the sound, no wonder people like this amp for bass, a good bass amp will usually have an oversized pwr trans,
not much varnish so lams came off easily,
this guy runs at about 12 kG at 60 CPS,
a 2 inch stack of 175 EI was used, 110 lams = 0.018" instead of the standard .014"
Lap 1 instead of the usual Lap 3,
low DCR means a stiff power source, no sag, less compressed sound, power amps modulate the pwr supply, so pwr transformers can influence the sound, no wonder people like this amp for bass, a good bass amp will usually have an oversized pwr trans,
not much varnish so lams came off easily,
Attachments
> out comes big bertha and in comes the electrical bill
At least it is Autumn.
Hottest day in July I was using my 300W Beauty to bus-up a tractor fuse-box.
> 5.94 V-ac from our test wind of 9 turns
6V/9t = 0.66V/turn (BIG transformer!)
> heater wind has 10 turns, ....6.58 V-ac
Cross-checked. (Nice when that happens.)
> 399 AC, {drop to} 355 V-ac sec volts
399V-355V = 44 Volts to lose.
44V/0.66 = 66 turns to come off.
> 67 Turns
Same answer.
> if removing turns is B-max, flux will go up
The Primary Volts and Primary turns are fixed quantities. Flux will not go up.
(If you took turns off the primary, but kept the 119V wall, then flux would go up.)
(If you added turns to the primary to get reduced secondary voltages everywhere, still at 119V input, flux would go down. But the only sane reason to do that is to re-wind a 110V primary for modern 125V walls. In that case you are just getting back to the original 1953 flux.)
> but there are losses
Yes, it will not come out the same as the math. Lower DCR suggests the voltage will be higher. However you are also re-loading the final for more current, so it could be lower. For 11% change, I would not expect 5% error, and 5% is "nothing" in tube-work. And the way you are doing it (un-winding the outside coil), you could just take-off your 60 turns, hook it up, beat it, and see where it came out. If distressingly high, take off 5 or 10 more turns to taste.
At least it is Autumn.
Hottest day in July I was using my 300W Beauty to bus-up a tractor fuse-box.
> 5.94 V-ac from our test wind of 9 turns
6V/9t = 0.66V/turn (BIG transformer!)
> heater wind has 10 turns, ....6.58 V-ac
Cross-checked. (Nice when that happens.)
> 399 AC, {drop to} 355 V-ac sec volts
399V-355V = 44 Volts to lose.
44V/0.66 = 66 turns to come off.
> 67 Turns
Same answer.
> if removing turns is B-max, flux will go up
The Primary Volts and Primary turns are fixed quantities. Flux will not go up.
(If you took turns off the primary, but kept the 119V wall, then flux would go up.)
(If you added turns to the primary to get reduced secondary voltages everywhere, still at 119V input, flux would go down. But the only sane reason to do that is to re-wind a 110V primary for modern 125V walls. In that case you are just getting back to the original 1953 flux.)
> but there are losses
Yes, it will not come out the same as the math. Lower DCR suggests the voltage will be higher. However you are also re-loading the final for more current, so it could be lower. For 11% change, I would not expect 5% error, and 5% is "nothing" in tube-work. And the way you are doing it (un-winding the outside coil), you could just take-off your 60 turns, hook it up, beat it, and see where it came out. If distressingly high, take off 5 or 10 more turns to taste.
Dreams
Well-known member
- Joined
- Jun 26, 2010
- Messages
- 331
Is the primary usually wound closest to the core?
If one were to choose to add windings to the primary, should it be right on top of the original primary (that for now, let's assume is closest to the core), or would adding windings on top of the outer windings be acceptable?
Would there be....what would the downsides to adding some outer primary windings be?
If one were to choose to add windings to the primary, should it be right on top of the original primary (that for now, let's assume is closest to the core), or would adding windings on top of the outer windings be acceptable?
Would there be....what would the downsides to adding some outer primary windings be?
i get nervous working with those pri winds as they sit on the power line, but hey, that's what fuses are for,
yes, usually the pri sits next to the core,
yes, usually the pri sits next to the core,
> would adding windings on top of the outer windings be acceptable?
Just throw it through the window. Anywhere.
I once did some semi-careful tests to learn the effect of a "part turn". This is for power transformers (wide-range audio iron is different).
Take fer example CJ's big V4 PT which he counts as 0.66 Volts per turn.
If you lay a wire next to the winding you get nearly nothing.
If you pass a wire _through_ one window of the core you get 0.33V.
That's the same whether it is a "1/8 turn" or a "7/8 turn".
If you pass a wire through _both_ windows of the core you get 0.66V.
That's the same whether it is a "1/2 turn" or a "11/8 turn".
What counts is the number of "window passings". Wire not in the window doesn't add voltage. Anywhere through the window counts the same.
So while the math is formulated as "turns", what really counts in iron-core work is "window passings". In the case of the V4 PT, 0.33V per passing (which is 0.66V per two passings which is the average number of passings per turn on a many-turn winding.)
This also suggests that we "can't" get closer to an exact voltage than 0.33V (for this PT). 10 turns (20 passings) is 6.6V. 10-1/2 turns (21 passings) is 6.93V. 0.33V increments. In fact, if we had to, we could slightly alter the primary turns, to get a different volts per "turn", and hit a specific target. In practice, we never need to get so very close.
In theory the "other" wire, or location within the window, matters.... but only about 0.1%. The magnetic effect is concentrated at the windows, because iron carries magnetism about 1,000 times better than anything else, so anything else is too small to matter. 6.3V or 6.3063V... who cares? Our "120V" is never 120.00V anyway.
Just throw it through the window. Anywhere.
I once did some semi-careful tests to learn the effect of a "part turn". This is for power transformers (wide-range audio iron is different).
Take fer example CJ's big V4 PT which he counts as 0.66 Volts per turn.
If you lay a wire next to the winding you get nearly nothing.
If you pass a wire _through_ one window of the core you get 0.33V.
That's the same whether it is a "1/8 turn" or a "7/8 turn".
If you pass a wire through _both_ windows of the core you get 0.66V.
That's the same whether it is a "1/2 turn" or a "11/8 turn".
What counts is the number of "window passings". Wire not in the window doesn't add voltage. Anywhere through the window counts the same.
So while the math is formulated as "turns", what really counts in iron-core work is "window passings". In the case of the V4 PT, 0.33V per passing (which is 0.66V per two passings which is the average number of passings per turn on a many-turn winding.)
This also suggests that we "can't" get closer to an exact voltage than 0.33V (for this PT). 10 turns (20 passings) is 6.6V. 10-1/2 turns (21 passings) is 6.93V. 0.33V increments. In fact, if we had to, we could slightly alter the primary turns, to get a different volts per "turn", and hit a specific target. In practice, we never need to get so very close.
In theory the "other" wire, or location within the window, matters.... but only about 0.1%. The magnetic effect is concentrated at the windows, because iron carries magnetism about 1,000 times better than anything else, so anything else is too small to matter. 6.3V or 6.3063V... who cares? Our "120V" is never 120.00V anyway.
> adding windings on top of ...windings
In blank-paper design:
The primary winding resistance affects the effective resistance of *all* other windings. If you have multiple secondaries, you want low primary resistance. Which means shortest length of turn. Which means the primary goes first, where the turns are short.
Your low-V high-I windings will be very fat wire. This does not want to bend around the sharp corners of the core. Its lumpiness is awkward to wind-over with smaller gauge. So "generally" the fat wire goes on the outside (though because this is longest length per turn and highest resistance, you might try to get it down inside).
The 600v winding is fairly accomodating. Fine wire wraps easily; in stuff smaller than a V4 PT it may come out "too fine" and you pick a size up from necessary so it doesn't break so often.
In retro-hacking, again, none of this matters as much as the labor to un-wind. You put your add-on over top of what you got.
The obvious downside is that if the PT was designed with a very sharp pencil, there will not be room to squeeze another layer without scraping on the window edge and shorting-out. Usually you also want to add outside insulation (paper/mylar) equivalent to the factory wrap so you don't get shocked through a pinhole in the wire varnish.
In blank-paper design:
The primary winding resistance affects the effective resistance of *all* other windings. If you have multiple secondaries, you want low primary resistance. Which means shortest length of turn. Which means the primary goes first, where the turns are short.
Your low-V high-I windings will be very fat wire. This does not want to bend around the sharp corners of the core. Its lumpiness is awkward to wind-over with smaller gauge. So "generally" the fat wire goes on the outside (though because this is longest length per turn and highest resistance, you might try to get it down inside).
The 600v winding is fairly accomodating. Fine wire wraps easily; in stuff smaller than a V4 PT it may come out "too fine" and you pick a size up from necessary so it doesn't break so often.
In retro-hacking, again, none of this matters as much as the labor to un-wind. You put your add-on over top of what you got.
The obvious downside is that if the PT was designed with a very sharp pencil, there will not be room to squeeze another layer without scraping on the window edge and shorting-out. Usually you also want to add outside insulation (paper/mylar) equivalent to the factory wrap so you don't get shocked through a pinhole in the wire varnish.
ok we get 352 V-ac sec which changes into 470 DC for the B+ with all 4 tubes plugged in,
bias on the two 270 ohm 10 watt cathode resistors which are in parallel is 32.8 V.
so 270/2 = 135 ohms, 32.8/135=242 ma plate current, so 60 ma per tube,
might hsave to bump the cathode resistance up a bit,
bias on the two 270 ohm 10 watt cathode resistors which are in parallel is 32.8 V.
so 270/2 = 135 ohms, 32.8/135=242 ma plate current, so 60 ma per tube,
might hsave to bump the cathode resistance up a bit,
> bump the cathode resistance
Well, the "470V B+" is just 437V plate-cathode, which at 0.060A is 26.22W plate+screen, probably 25W plate alone.
Well under the nominal "30W" for TV-tech 6L6. About 83%.
The 70% guide is for fix-biased amps which can rise with signal, but must be aimed a little cool to cover bias drift.
A cathode-biased stage "should" be fine at 99.44% of rating.
Personally I hit a run of 6L6 (old and semi-new) which were very drifty. I do not think it wise to run these anywhere near the rating.
This in contrast to several 6550 (very old and very new) which biased rock-solid, and I fearlessly let them idle at 41W Pdiss.
I think the big Ampegs *should* run 6550.
I think it will be cheaper over a decade.
Quick turn-around jobs, you do what works.
Well, the "470V B+" is just 437V plate-cathode, which at 0.060A is 26.22W plate+screen, probably 25W plate alone.
Well under the nominal "30W" for TV-tech 6L6. About 83%.
The 70% guide is for fix-biased amps which can rise with signal, but must be aimed a little cool to cover bias drift.
A cathode-biased stage "should" be fine at 99.44% of rating.
Personally I hit a run of 6L6 (old and semi-new) which were very drifty. I do not think it wise to run these anywhere near the rating.
This in contrast to several 6550 (very old and very new) which biased rock-solid, and I fearlessly let them idle at 41W Pdiss.
I think the big Ampegs *should* run 6550.
I think it will be cheaper over a decade.
Quick turn-around jobs, you do what works.
here is a schemo of the final product,
screen resistors seem to make a difference tone wise,
you can hear the difference between a set of 2.2K resistors and 3.9k resistors,
3.9K sounds slightly better, seems to slow down the attack time of the notes,
cathode resistors were changed from two 270 ohm/10 watt in parallel to two 430 ohm 25 watt in parallel, so we go from 135 ohm 20 watt to 215 ohm 50 watt,
215 ohms with the 3.9K screen resistors and the reduced plated voltage of 470 volts sounds way cool, like an old Bassman, instead of the beaststron V4 ear banger, hope the guy likes it,
with a full signal crankin, the 50 watts of cathode resistors still get warm, so you can never have too much safety factor for resistors in this location,
screen resistors seem to make a difference tone wise,
you can hear the difference between a set of 2.2K resistors and 3.9k resistors,
3.9K sounds slightly better, seems to slow down the attack time of the notes,
cathode resistors were changed from two 270 ohm/10 watt in parallel to two 430 ohm 25 watt in parallel, so we go from 135 ohm 20 watt to 215 ohm 50 watt,
215 ohms with the 3.9K screen resistors and the reduced plated voltage of 470 volts sounds way cool, like an old Bassman, instead of the beaststron V4 ear banger, hope the guy likes it,
with a full signal crankin, the 50 watts of cathode resistors still get warm, so you can never have too much safety factor for resistors in this location,
Attachments
three unique aspects of this amp are
red circle-diode in ground buss which raises preamp grounds by 0.6 V-dc (forward drop of diode) this is probably to get the ground off the noise floor, a trick seen in some old HP volt meters,
green circle-standby switch is in the ground leg instead of the B+, it seems to operate in a smooth fashion without any snaps and pops,
blue circle-instead of a secondary tap for the fixed bias circuit, a cap is used as an AC resistor to drop the plate voltage down to a reasonable level without heat,
but we are not using this circuit as the amp now runs in cathode bias,
red circle-diode in ground buss which raises preamp grounds by 0.6 V-dc (forward drop of diode) this is probably to get the ground off the noise floor, a trick seen in some old HP volt meters,
green circle-standby switch is in the ground leg instead of the B+, it seems to operate in a smooth fashion without any snaps and pops,
blue circle-instead of a secondary tap for the fixed bias circuit, a cap is used as an AC resistor to drop the plate voltage down to a reasonable level without heat,
but we are not using this circuit as the amp now runs in cathode bias,
Attachments
there is a tropical fish capacitor that has a cracked exterior, might have been fine but we replaced it just in case, did not hear a big drop in sound quality, but some folks claim that these caps have magical properties, mostly the stomp box crowd,
Magnavox tube means it is original, voltages checked out ok so we left it in there,
.33 uf coupling caps mean plenty of bass, so this amp gets used for bass and keyboards also,
pull two of the 7027 pwr tubes if you have a stock V4 and you will have a better match with the 2000 turn OPT pri. this will make the sound guy happy also,
Magnavox tube means it is original, voltages checked out ok so we left it in there,
.33 uf coupling caps mean plenty of bass, so this amp gets used for bass and keyboards also,
pull two of the 7027 pwr tubes if you have a stock V4 and you will have a better match with the 2000 turn OPT pri. this will make the sound guy happy also,
Attachments
I have 4 of them but just one was fine until damage pin the filter power section fail!!.....all of them wait for repair in my shelf . I compared to the SVT amp and 300 PS of Fender with bass and its sound was very clear tono than 300 PS. And SVT too but with less power able than SVT but more than 300 PS.
Opacheco.
Opacheco.
CJ said:
