GREETINGS ALL, I have a question regarding connecting output transformers in series. I am in the process of building an 833A SE amp. I have searched for a suitable output transformer at a reasonable price. Most seem to be about £750 for a custom made one with a 10K primary. So here is my question. If I buy say 2 Hammond transformers with 5k primaries and connect them in series (and connect the 2 x 4 ohm secondary taps also in series) would that give me a 10k primary into 8 ohms. If so what would the downside be and would it even work
The turn ratio will be the same. So impedance seen at the anode will be
the same as with only one transformer.
Rolf
the same as with only one transformer.
Rolf
I believe you are referring to the case where the windings are all on the same core, like how some power transformers are made, then yes. In this case, it is two separate SE transformers, i.e., they have separate cores.
Maybe a better way is to use a single transformer but connect the load to lower sec tap, as primary impedance also depends on the value of secondary load. If you connect exactly 8 ohms to sec than reflected impedance is 5K on primary and 16 ohms load reflected impedance is 10K. So you have to connect 8 ohms speaker to 4 ohms tap make it 10K.
OUTPUT TRANSFORMERS IN SERIES
Gentlemen, thank you for the response thus far. I am of a mind to think my suggestion may have some merit....but the proof of the pudding would be in the eating of course. If I had a couple of SE transformers knocking about I would give it a go (even if its connected to something simple like a 6BQ5) it is either going to work or its not. If it does work then just maybe we might have discovered a whole new and cheaper way and maybe even better way of doing things......I would be delighted if someone could try it out even on an existing amplifier that way it would give an objective test.
Gentlemen, thank you for the response thus far. I am of a mind to think my suggestion may have some merit....but the proof of the pudding would be in the eating of course. If I had a couple of SE transformers knocking about I would give it a go (even if its connected to something simple like a 6BQ5) it is either going to work or its not. If it does work then just maybe we might have discovered a whole new and cheaper way and maybe even better way of doing things......I would be delighted if someone could try it out even on an existing amplifier that way it would give an objective test.
I have connected OPT's together in many different ways, and most work in most situations.
Two SE OPT's can be connected together with the primaries in series and the secondaries in parallel. If using a Hammond OPT with multiple taps, wire all of them in parallel and then choose the tap that works best with your tube, speaker and B+ combination.
The 833A is a rather hard tube to get an OPT for because of its high plate resistance. I had a well known OPT builder make me an OPT specifically for the 833A several years ago and there were problems with the high frequency response due to leakage inductance.
I used a 5K load with 1500 volts to get about 200 watts from the 833A. Forum member MAGZ used a 10K load with 2300 volts to get about the same power.
Hammond SE PTs are known for their resonant notch I the high frequency response. The notch is in the 17 to 25 KHz range and it will move with the parameters of the tube driving them. Connecting two of them together will produce unpredictable results.
I have a pair of 1628SEA's in storage somewhere, but I have recently moved and am now in the process of building my new lab and haven't got a way to test them right now.
Two SE OPT's can be connected together with the primaries in series and the secondaries in parallel. If using a Hammond OPT with multiple taps, wire all of them in parallel and then choose the tap that works best with your tube, speaker and B+ combination.
The 833A is a rather hard tube to get an OPT for because of its high plate resistance. I had a well known OPT builder make me an OPT specifically for the 833A several years ago and there were problems with the high frequency response due to leakage inductance.
I used a 5K load with 1500 volts to get about 200 watts from the 833A. Forum member MAGZ used a 10K load with 2300 volts to get about the same power.
Hammond SE PTs are known for their resonant notch I the high frequency response. The notch is in the 17 to 25 KHz range and it will move with the parameters of the tube driving them. Connecting two of them together will produce unpredictable results.
I have a pair of 1628SEA's in storage somewhere, but I have recently moved and am now in the process of building my new lab and haven't got a way to test them right now.
Assuming I've grasped this properly, I think it would work with proviso that its recognised that dc flows in each primary, and the tx can tolerate that without saturation.
Nice, I'll tuck this away for a rainy day.
Last edited:
Ideally, one can connect the OTs up just like resistors (the windings not being on the same core), but common mode capacitance will be a new issue. (and DC current specs still need to be obeyed in SE mode)
The OT is normally wound to minimize distributed capacitance by keeping the AC grounded end (B+) of the primary nearest to the secondary or core. With two OTs, one OT primary will have the whole primary winding swinging around a 1000 volts or so.
Compared to one OT, the pair will require more net primary turns to meet the primary inductance requirement. This increases the distributed capacitance further. (and leakage L) So performance will be worse with combined OTs versus a single made to spec OT. Putting two resonant circuits together (leakage L and distr. C) may cause a lower resonance besides (with 2L and 2C, f = 2pi/Sqrt(4LC) ) .
Net Leakage L might stay the same, since they get effectively paralled, then transformed to 2X Z by series connection at the output. Distributed C will go up more than 2X with the common mode component added though.
If Tubelab (George) had problems with a custom made single OT for the 833A, a doubled up one will likely have bigger issues with 833A at 10K. I recall Wavebourne mentioned some problems with a custom made 10K Edcor OT for GU50s.
The OT is normally wound to minimize distributed capacitance by keeping the AC grounded end (B+) of the primary nearest to the secondary or core. With two OTs, one OT primary will have the whole primary winding swinging around a 1000 volts or so.
Compared to one OT, the pair will require more net primary turns to meet the primary inductance requirement. This increases the distributed capacitance further. (and leakage L) So performance will be worse with combined OTs versus a single made to spec OT. Putting two resonant circuits together (leakage L and distr. C) may cause a lower resonance besides (with 2L and 2C, f = 2pi/Sqrt(4LC) ) .
Net Leakage L might stay the same, since they get effectively paralled, then transformed to 2X Z by series connection at the output. Distributed C will go up more than 2X with the common mode component added though.
If Tubelab (George) had problems with a custom made single OT for the 833A, a doubled up one will likely have bigger issues with 833A at 10K. I recall Wavebourne mentioned some problems with a custom made 10K Edcor OT for GU50s.
Last edited:
Doesn't the SE swing around 2xEbb, with Magz's example, thats' ~4600V, so doesn't each OPT sees about 2300V? Also how do you calculate the turns ratio when the 2 OPTs' primary are connected in series but the cores are not magnetically coupled?
Treat them as completely independent.
I would be really concerned about whether the insulation in those transformers can withstand the voltages typically employed with the 833.
Consider also the effect of both additional leakage inductance and capacitance primary to core.
I think I would save my pennies for a suitable single OPT rather than wiring a couple of slightly cheaper transformers in series. (Possibly you are not really saving any money over a carefully chosen EU vendor)
I have not done business with these guys, but they seem to be well regarded here: Ogonowski.eu - profesjonalne transformatry do wzmacniaczy lampowych They make custom SE output transformers to order and the site offers an English language option.
http://www.ogonowski.eu/transformers/single-ended/
I would be really concerned about whether the insulation in those transformers can withstand the voltages typically employed with the 833.
Consider also the effect of both additional leakage inductance and capacitance primary to core.
I think I would save my pennies for a suitable single OPT rather than wiring a couple of slightly cheaper transformers in series. (Possibly you are not really saving any money over a carefully chosen EU vendor)
I have not done business with these guys, but they seem to be well regarded here: Ogonowski.eu - profesjonalne transformatry do wzmacniaczy lampowych They make custom SE output transformers to order and the site offers an English language option.
http://www.ogonowski.eu/transformers/single-ended/
The last 3 posts mention several very valid concerns, and yes the voltage swing across an SE OPT is 2 X the B+ voltage under NORMAL operating conditions. I have seen more than 3 X B+ when driving an SE amp into hard clipping with a guitar preamp into a real speaker load.
The capacitance and inductance effects will be different than the usual single OPT. There could be multiple notches in the HF response due to interactions between the leakage inductance of one OPT and the capacitance to ground in the other OPT.
All of my experiments have been with small inexpensive OPT's, or push pull OPT's that I have several of.
The capacitance and inductance effects will be different than the usual single OPT. There could be multiple notches in the HF response due to interactions between the leakage inductance of one OPT and the capacitance to ground in the other OPT.
All of my experiments have been with small inexpensive OPT's, or push pull OPT's that I have several of.
- Status
- Not open for further replies.