Hi, after building two triode SE amps, I decided to build a small pentode SE amp. One of the reasons is to learn more about pentodes (when not used as triodes) since I learn better when I have a project ongoing.
I bought a couple of TGL5/002 transformers because they were inexpenive, small size and readily available for me. I also used TGL20/003 transformers in another amp (both 6AS7 triodes in parallel) and that works very well, at least for my ears.
However, once I built a test circuit I found that these transformers have a problem with low frequencies. The specifications say that the frequency range is 40Hz - 20kHz and power is 5W, however, I get these results, when using different tubes ( the same global NFB resistor was used).
6P3S/6L6 in pentode mode - low frequencies start to drop off at around 300Hz, with -3dB being somewhere around 200Hz. Up to about 4W with 2% THD (with global NFB). The frequency response stays about the same on different output levels. I tried it without NFB (just the output stage) and got very similar results.
6P14P/EL84 in pentode mode - low frequencies do not drop off as much (-3dB at probably 70Hz), but low frequency signals get distorted a lot (see attachment) if I try to get out more than about 0.25W from it. On 1kHz, I get almost 4W.
6P3S/6L6 in triode mode - lower power overall, but better low frequency reponse than EL84 in pentode mode and less distortion.
Does this mean that the transformer cannot work at low frequencies, or is it just that I have not figured out how to make it work? I have not seen that type of distortion before.
I bought a couple of TGL5/002 transformers because they were inexpenive, small size and readily available for me. I also used TGL20/003 transformers in another amp (both 6AS7 triodes in parallel) and that works very well, at least for my ears.
However, once I built a test circuit I found that these transformers have a problem with low frequencies. The specifications say that the frequency range is 40Hz - 20kHz and power is 5W, however, I get these results, when using different tubes ( the same global NFB resistor was used).
6P3S/6L6 in pentode mode - low frequencies start to drop off at around 300Hz, with -3dB being somewhere around 200Hz. Up to about 4W with 2% THD (with global NFB). The frequency response stays about the same on different output levels. I tried it without NFB (just the output stage) and got very similar results.
6P14P/EL84 in pentode mode - low frequencies do not drop off as much (-3dB at probably 70Hz), but low frequency signals get distorted a lot (see attachment) if I try to get out more than about 0.25W from it. On 1kHz, I get almost 4W.
6P3S/6L6 in triode mode - lower power overall, but better low frequency reponse than EL84 in pentode mode and less distortion.
Does this mean that the transformer cannot work at low frequencies, or is it just that I have not figured out how to make it work? I have not seen that type of distortion before.
Last edited:
First of all, are these transformers made for single ended amplifiers (are the laminations with all the Es on one side, and all the Is on the other sides? How many mA DC are the transformers rated for? What are the transformer's specifications? How much quiescent current does the output tube have?
I appears that there is not a lot of inductance in the primary windings, or enough DC current ability, or not enough laminations. Without negative feedback, there will be distortion and low frequency roll off.
The lower the impedance of the tubes (rp) driving the transformer, the less low frequency distortion and low frequency roll off there will be. A pentode will have high impedance, the same pentode that is connected in triode mode will have lower impedance.
Negative feedback can reduce the problem of low frequency roll off, and low frequency distortion; but it can not completely eliminate it if there is not enough inductance, and not enough DC current rating, and not enough laminations.
How about a schematic of the amplifier, and some voltage and current values, resistance values, etc.
I appears that there is not a lot of inductance in the primary windings, or enough DC current ability, or not enough laminations. Without negative feedback, there will be distortion and low frequency roll off.
The lower the impedance of the tubes (rp) driving the transformer, the less low frequency distortion and low frequency roll off there will be. A pentode will have high impedance, the same pentode that is connected in triode mode will have lower impedance.
Negative feedback can reduce the problem of low frequency roll off, and low frequency distortion; but it can not completely eliminate it if there is not enough inductance, and not enough DC current rating, and not enough laminations.
How about a schematic of the amplifier, and some voltage and current values, resistance values, etc.
Last edited:
The specs are a vague but it's not surprising it is bad with 6L6. It's a tiny transformer. I guess you are running too high DC current with both 6L6 and EL84 in pentode mode. Although at 1KHz it might be ok (because the AC induction required to pass 4W is very small) as soon as you go low in frequency it saturates! Saturation will destroy inductance and so frequency response will be bad even at 0.25W.....
A bandwidth of 40Hz-20KHz means that these are the -3 dB points. I would guess that optimal DC current you can run at is likely around 25 mA. With such DC current possibly it will work as a 5VA transformer. 5VA means 5W at 50Hz which in turns means 1.8W at 30Hz. In other words it's a 2W full bandwidth transformer and you should make the amp accordingly. All this ASSUMING it is a GAPPED transformer. Otherwise it is not suitable for SE....
A bandwidth of 40Hz-20KHz means that these are the -3 dB points. I would guess that optimal DC current you can run at is likely around 25 mA. With such DC current possibly it will work as a 5VA transformer. 5VA means 5W at 50Hz which in turns means 1.8W at 30Hz. In other words it's a 2W full bandwidth transformer and you should make the amp accordingly. All this ASSUMING it is a GAPPED transformer. Otherwise it is not suitable for SE....
Last edited:
TGL series technical data from Indel web site: Katalog A
The TGL5/002 is a 5W transformer for EL84 single ended at 48mA. It is a comparatively big EL84 transformer (1 Kg weight) and I was considering it for a future build. The bad low frequency response you documented seems strange. Is your speaker impedence 4 ohms?
The TGL5/002 is a 5W transformer for EL84 single ended at 48mA. It is a comparatively big EL84 transformer (1 Kg weight) and I was considering it for a future build. The bad low frequency response you documented seems strange. Is your speaker impedence 4 ohms?
Pcan
1Kg transformer IS a tiny transformer, especially if the core is EI type (because it is less efficient than C or toroidals) Half of the weigh is possibly just copper and other stuff. 48 mA is the max DC you can apply and it does not specify much. It's not clear if this is a max DC because of the copper wire size, saturation or operative. Because the transformer is tiny and the results above in practical amps are not up to expectations, I think 48 mA is not the max DC operative current. 25-30 mA is likely right.
1Kg transformer IS a tiny transformer, especially if the core is EI type (because it is less efficient than C or toroidals) Half of the weigh is possibly just copper and other stuff. 48 mA is the max DC you can apply and it does not specify much. It's not clear if this is a max DC because of the copper wire size, saturation or operative. Because the transformer is tiny and the results above in practical amps are not up to expectations, I think 48 mA is not the max DC operative current. 25-30 mA is likely right.
I'm not familiar with these transformers. I do however have experience with smallish SE transformers from Triode Electronic (tiny), Edcor and Hammond. They have all worked fine within their limitations. I think it is fair to assume the TGL5/002 will too in the right circuit. That said, it's difficult to come up with good advise without a schematic and pictures of the test rig. As others have pointed out (before diverting into peripheral subjects), first thing is to confirm that max idle current is not exceeded and the load is correct.
Don't know TGL5/002 but I had a look at TGL5/001:
TGL5/001 is supposed to be SE transformer 5VA.
EI78 (1.2kg, 2.5 pounds)
Only one interleave prim / sec (sec in middle)
Primary inductance at 50Hz (simple V/I method) a mere 4.3 Henry
DCR primary 30 ohm --> very low
they use 0.315mm AWG28 primary wire --> big for 50mA DC --> low number of turns
It looks as if they took one of their 20VA Push-Pull, same core and wire, re-assembled with non-interleaved iron, added air-gap, 0.06mm foil.
And indeed - even the primary center tap is still present !!!
TGL5/001 is supposed to be SE transformer 5VA.
EI78 (1.2kg, 2.5 pounds)
Only one interleave prim / sec (sec in middle)
Primary inductance at 50Hz (simple V/I method) a mere 4.3 Henry
DCR primary 30 ohm --> very low
they use 0.315mm AWG28 primary wire --> big for 50mA DC --> low number of turns
It looks as if they took one of their 20VA Push-Pull, same core and wire, re-assembled with non-interleaved iron, added air-gap, 0.06mm foil.
And indeed - even the primary center tap is still present !!!
I too investigated the Indel TGL5/001, which appears to be a version of the aforementioned OPT with an 8ohm secondary, and to be honest, I wasn't impressed. Below 100Hz I didn't get much more than a quarter of a watt before I could notice the distortion at the screen of an oscilloscope. And when it can be seen on a scope, it's more than audible as a rule of thumb.
This was on an amp with an ECC81 talking to an EL84 in triode mode I designed as a kit to be beginner accessible for a local publication. So I very quickly gave up this Polish transformer for a Canadian alternative. I tried a few transformers that could be obtained from local sources without resorting to ebay-shopping, and the similarly sized Hammond 125DSE was performing miles beyond this Indel tranny.
Now, I don't know whether the bad low end performance was due to poor air gapping of the core, or something else, since after seeing the bass performance I dismissed these trannies almost at once. Maybe I'll one day do some more testing, maybe trying to vary the air gap, just for fun.
From left to right the image shows the scope trace for Indel TGL5/001, Hammond 125DSE and Hammond 125ESE at 1W to 8ohms at 50Hz. Need I say any more?
This was on an amp with an ECC81 talking to an EL84 in triode mode I designed as a kit to be beginner accessible for a local publication. So I very quickly gave up this Polish transformer for a Canadian alternative. I tried a few transformers that could be obtained from local sources without resorting to ebay-shopping, and the similarly sized Hammond 125DSE was performing miles beyond this Indel tranny.
Now, I don't know whether the bad low end performance was due to poor air gapping of the core, or something else, since after seeing the bass performance I dismissed these trannies almost at once. Maybe I'll one day do some more testing, maybe trying to vary the air gap, just for fun.
An externally hosted image should be here but it was not working when we last tested it.
From left to right the image shows the scope trace for Indel TGL5/001, Hammond 125DSE and Hammond 125ESE at 1W to 8ohms at 50Hz. Need I say any more?
4.3H is really poor. 0.315 mm is without a doubt exaggerated for that core and won't allow enough turns. It is NOT 5VA by any standard. With 4.3 and 48 mA DC the max undistorted primary voltage you can apply at 50Hz is barely 45V rms which means less than 0.4W! At 100Hz it becomes 1.5W into nominal load.
For my PCL86 triode strapped back-up amp that I have on my bench to listen to the radio I have made a 7K OPT using a smaller EI-75 core 2.5x2.5 cm cross- section with 0.4 dB power loss (91% efficiency) and 42H inductance at full throttle. It has 3024 primary turns using 0.16 mm wire (0.2mm considering insulation). Output power at clipping is a just over 2W down to 35Hz. I run the PCL86 with fixed bias at 340V plate voltage though, which is a bit out of spec but in 7 years I have never had a problem. The valves are Yugoslavian made in the 70's....
From what "playloadde" measured the max output power at 50Hz is just under....0.4W!!!
Incredible how people can sell such garbage. It's not remotely close to specs. I wonder what 5 VA might mean.
Thanks for the help, I will try to address all your replies:
I measured the frequency response etc with a 4.1Ohm (2x8.2Ohm in parallel) resistor attached to the output. I do not have a circuit diagram, but will write the current and voltage values for my test circuits:
EL84 pentode:
B+ 309V
G2 220V (From B+ rough a 23K resistor, with a 47u cap)
Cathode voltage 5.1V
Cathode reesistor 120Ohm
This should result in 3.1mA G2 current and 41mA plate current.
Low frequencies get distorted above 0.25W, but they do not roll off. If I play a song that has a lot of bass in it, the bass gets distorted if I turn up the volume.
6L6 Pentode:
B+ 310V
G2 260V
Cathode reistor 410Ohm
Cathode voltage 18.4V
This gives 2.1mA for G2 and 42mA for plate.
Low frequencies drop off a lot, but do not get distorted as much (probably because they drop off low enough). If I play a song, it song as if the bass has been filtered out. The frequency response does ot change when I turn up the volume.
6L6 triode:
B+ 296V
Cathode resistor 333Ohm
Cathode voltage 19.4V
This results in 58mA trough the plate, so above the rating of the transformer. However now, the low frequencies do not drop off and I can get at least 1W at 70Hz. I currently do not know the maximum power because the input stage does not have enough swing.
The transformer has a thin air gap. It has a tap on the primary, it is about center, so I guess I could also use this transformer for PP, though I do not know if the air gap would cause problems.
TGL5/002 is the same as TGL5/001, but the secondary on 002 is 4Ohm, where the secondary on 001 is 8Ohm.
It looks like 6P41S can also work with 5K load. I'll try to use that and see what happens. If it doesn't work right, I'll pull out the Hammond 125ESE (i have it in a box somewhere) and test with that.
I measured the frequency response etc with a 4.1Ohm (2x8.2Ohm in parallel) resistor attached to the output. I do not have a circuit diagram, but will write the current and voltage values for my test circuits:
EL84 pentode:
B+ 309V
G2 220V (From B+ rough a 23K resistor, with a 47u cap)
Cathode voltage 5.1V
Cathode reesistor 120Ohm
This should result in 3.1mA G2 current and 41mA plate current.
Low frequencies get distorted above 0.25W, but they do not roll off. If I play a song that has a lot of bass in it, the bass gets distorted if I turn up the volume.
6L6 Pentode:
B+ 310V
G2 260V
Cathode reistor 410Ohm
Cathode voltage 18.4V
This gives 2.1mA for G2 and 42mA for plate.
Low frequencies drop off a lot, but do not get distorted as much (probably because they drop off low enough). If I play a song, it song as if the bass has been filtered out. The frequency response does ot change when I turn up the volume.
6L6 triode:
B+ 296V
Cathode resistor 333Ohm
Cathode voltage 19.4V
This results in 58mA trough the plate, so above the rating of the transformer. However now, the low frequencies do not drop off and I can get at least 1W at 70Hz. I currently do not know the maximum power because the input stage does not have enough swing.
The transformer has a thin air gap. It has a tap on the primary, it is about center, so I guess I could also use this transformer for PP, though I do not know if the air gap would cause problems.
TGL5/002 is the same as TGL5/001, but the secondary on 002 is 4Ohm, where the secondary on 001 is 8Ohm.
It looks like 6P41S can also work with 5K load. I'll try to use that and see what happens. If it doesn't work right, I'll pull out the Hammond 125ESE (i have it in a box somewhere) and test with that.
Pentium if you change the tube type power handling won't be different. It's a waste of time as well. It looks like you have the opposite problem I thought in the beginning: because the inductance is too low the tube soon runs out of current as frequency goes down. With the 6L6 in triode you get 1W just because the current is higher. 1W output for 16W plate dissipation!!!
I suggest that you use the Hammond and forget about this "thing". I just can't classify it as output transformer. Maybe it's good for telephony but not for HiFi....
I suggest that you use the Hammond and forget about this "thing". I just can't classify it as output transformer. Maybe it's good for telephony but not for HiFi....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.