What methods and techniques are there for protecting tubes and the output transformers in the case of a short or open circuit? I am building a tube amplifier using kt88's in a push-pull configuration. I would like the amplifier to have protection incase the amplifier is ran without a load or if a load shorts.
Thanks!
Thanks!
Fuses! Thats a simple and cheap way, thanks. I have very little experience with tube amps and this is my first build. I am assuming fuses will protect against short circuits but what about open circuits? I have read that having no load connected to a tube amplifier can damage that output transformer. Is there anyway of protecting from an open circuit?
A Tube amp with Transormer output will be damaged by VOLTAGE generated in the O/P Tx as there is no load to control it.
Fuses Will NOT protect for this condition--Until its TOO Late!
A 'Snubber' network on the Primary of summit like a 0.01 and 100 ohm can help prevent excess volts from destroying the trans. if its run no-load....
NOT IN THIS CASE! A tube amp, especially a pentode amp can be damaged if the load goes open during use. An open load draws NO current. In normal operation one (or more) output tube is conducting a fixed current into the OPT primary. The OPT will transform the primary current into a voltage that is applied to the load. If the load is suddenly removed the current has nowhere to go so the voltage will rise to an extremely high value (several KV is possible) causing an arc somewhere. The situation is magnified several times over if the amp is operating at the clipping level (a guitar amp) since the output tubes are now functioning like switches and the OPT works like an ignition coil.
The most likely place for an arc to start is inside the OPT itself. This will fry it quickly. Most of the common audio output tubes have the plate on pin 3 and the heater on pin 2. I have seen an arc start at the base of one of the output tube sockets from pin 3 to pin 2. If the heater winding is not grounded it is possible for this to blow the power transformer. Very high power guitar amps (the Ampeg SVT) have been known to burst into flames if the speaker goes open during use. I saw it happen back in the early 70's. Autopsy revealed blown OPT, blown power transformer, two of the 6 6550's and several small parts, all fried because a 300 watt amp was played into a 100 watt speaker cabinet at an outdoor show where everything was set on 11.
There is no way to protect against this with a fuse. The usual protection method is to place a 100 ohm 10 watt or so resistor across the speaker terminals. This will steal a couple of watts of output power, and is not a guarantee either. Gas tube spark arresters like used in telephone equipment and big zener diodes across the speaker terminals is another possibility.
NEVER FUSE THE SPEAKER LEADS IN A TUBE AMP that uses an OPT. OTL's may need fuses to protect the speakers from blown tubes.
My prefered protection is:
Use a MOV and series resistor across each OT primary half winding. DC working voltage of MOV slightly greater than max DC of HT supply. Resistor = approx half PP loading impedance - used to damp the transient if MOV starts conducting. Use smallest disc MOV available to keep capacitance down to a negligible level (the 7mm disc types typically have low capacitance <100pF).
Use a fuse in each cathode, with a parallel high value cathode bias resistor (say 10k) to bias tube deep in to cutoff, but not exceed heater-cathode voltage rating. You can put a 1R current sensing resistor from 0V to fuse, for easy bias maintenance checking. Aim for fuse rating at or just above the peak tube current, to give nominal 20% above likely worstcase rms.
Ciao, Tim
Use a MOV and series resistor across each OT primary half winding. DC working voltage of MOV slightly greater than max DC of HT supply. Resistor = approx half PP loading impedance - used to damp the transient if MOV starts conducting. Use smallest disc MOV available to keep capacitance down to a negligible level (the 7mm disc types typically have low capacitance <100pF).
Use a fuse in each cathode, with a parallel high value cathode bias resistor (say 10k) to bias tube deep in to cutoff, but not exceed heater-cathode voltage rating. You can put a 1R current sensing resistor from 0V to fuse, for easy bias maintenance checking. Aim for fuse rating at or just above the peak tube current, to give nominal 20% above likely worstcase rms.
Ciao, Tim
In a PP amp the voltage of the anodes will swing to up to twice as high as the HT supply. An MOV will typically clamp at twice the rated working voltage, so something slightly greater than the HT for the rated working voltage should be good, but be very careful how you read the specs when you buy MOVs.
The fuse idea I'm not so fond of. For one thing, the parallel fuse/resistor idea on the cathode is a good way to burn up the other output tube if you're using a CCS on your output stage like I do - it will force the other tube to double its bias current. For another thing the fuse is highly non-linear as it heats and cools. I would just take my chances. If the fuse blows the tube is probably shot anyway. I'm not sure what you're protecting there. The amp fusing should kick in for something catastrophic.
As for a short in the output transformer, I feel safer with a large current capacity MOV on the secondary of the OT to ground. If the primary shorts internally to the secondary there's no guarantee where it will happen, and any fusing may or may not blow on the primary side depending on the location of the short.
Just my 2 cents.
PS: Why would you run your amp with no load???
Last edited:
In a PP amp the unloaded winding will nominally have a voltage equal to the turns ratio - ie. HT to a peak level determined by opposite loaded anode in saturation (a magnitude slightly less than HT). If there is blocking going on then both windings can be unloaded. If the output is unloaded, then things get uglier. The aim of the MOV is to start conducting if a substantial overvoltage peak occurs, and even then it is a damped loading applied to the overvoltage portion only. Yes the MOV specified DC operating level has to be somewhat above HT, and then there is a large tolerance on the 1mA DC voltage level, which is somewhat higher still. The intent is to alleviate overvoltage stress on OT insulation - as the OT is typically more expensive than the tubes, and sometimes there are no replacements per se.
I must admit I have not heard of OT's shorting primary to secondary, but it is not uncommon for instrument amps to have the speaker plug pulled out, or someone to change speakers during testing and to forget one terminal.
Agreed that the fuse and parallel resistor technique is for typical fixed bias PP. The intent is to provide a lower level of over-current stress to the OT (than say a HT fuse) - for same reasons as above. The fuse voltage drop, if sized for no more than 70-80% rated, is I suggest negligible (I've seen 0.11 ohm for a hot 2A 3AG) for most applications (especially instrument amps).
Ciao, Tim
I must admit I have not heard of OT's shorting primary to secondary, but it is not uncommon for instrument amps to have the speaker plug pulled out, or someone to change speakers during testing and to forget one terminal.
Agreed that the fuse and parallel resistor technique is for typical fixed bias PP. The intent is to provide a lower level of over-current stress to the OT (than say a HT fuse) - for same reasons as above. The fuse voltage drop, if sized for no more than 70-80% rated, is I suggest negligible (I've seen 0.11 ohm for a hot 2A 3AG) for most applications (especially instrument amps).
Ciao, Tim
definelty fuses. keep resistors slightly undersized compared to the tubes max power ratings so that the resisters blow first instead of the tubes or trannies. inrush limiters on the power supply. warm up circuit approaches. also think about what could be done in the driver tube area to prevent driver tube problems from feeding into the output stages. a shorted driver tube could glow red a power tube never blowing fuses.
Nobody here would. But if the dog wanders behind the gear and rips out the speaker leads, most of our spouses will happily listen to one channel of background music until hubby returns to fix the other. (...or until the smoke alarm triggers.)
...Todd
Thanks for all the replies. I will definatley look into using MOVs in order to prevent high voltages from damaging the output transformer.
For the short circuit protection, could I just use a current sense resistor (~ 0.1 Ohm) on the HT and then turn off the HT when the current gets too high? I will also use a fuse but I want this amp to be as robust as possible to any abuse. I have the power supply mostly designed and it includes a 20 sec delay and soft start for the HT voltage.
For the short circuit protection, could I just use a current sense resistor (~ 0.1 Ohm) on the HT and then turn off the HT when the current gets too high? I will also use a fuse but I want this amp to be as robust as possible to any abuse. I have the power supply mostly designed and it includes a 20 sec delay and soft start for the HT voltage.
A couple of my Conrad Johnson's (MV-50 and MV-55) fuse the B+ to the center tap of the OPT. On the MV-50 they use 2 - 270K, 2W resistors in parallel then an LED in series & this string is in parallel with a 1 amp, 600v fuse (BBS-1 or BLS-1). If a fuse blows the LED lights. Seems to work. I've blown a couple fuses when tubes went bad.
Your initial question was about how to protect against no load and shorts, I assume a shorted output.
The MOVs should protect the transformer from no load conditions (BTW, I was being sarcastic about why anyone would run an amp with no load having done it myself more than once
). Using fuses to protect the tubes at the possible expense of fidelity is not a wise move IMHO, but again, that's one person's opinion. I would try to keep all fusing on the mains side of the power tranny if possible, or certainly before any filtering on the secondary side of the power tranny. If you use good quality KT88s and you don't abuse the amp I think you're a little overly concerned, but it's your amp and your application. I think all the advice in this thread taken together should be a good starting point.
Fusing the B+ only to pentode PP output stage (eg. KT88) has the disadvantage that if the fuse goes then the tube still has screen supply. Fuse may go through age, or for an event that doesn't damage the tubes, but likely that tubes will be damaged with just screen supply energised. Hence a common preference for fusing the cathodes. Not sure how well the parallel resistor-LED alleviates the stress on the tube - but sounds reasonable for a triode PP.
Some prefer to use a small cathode sense resistor that will 'burn up' on overload, as a poor mans fuse - but that has a rather ill-defined I2t characteristic, and provides an ignition source, and doesn't protect the heater-cathode insulation if the tube is not at fault. Fuses are pretty quality consistent with respect to fusing I-t characteristic, and can be fairly well matched to a PP loadline.
The thermal mass of the fuse element is likely to attenuate the voltage drop parasitic signal that adds to the fixed bias level in a fixed bias PP stage. Given that a hot 2A fuse is circa 0.11V steady state, I would anticipate any thermally induced parasitic signal from resistance change would be lucky to get in to the 10's of mV, and that would only be likely at low frequencies. The parasitic effect is likely to be even less in a cathode biased PP.
Ciao, Tim
Some prefer to use a small cathode sense resistor that will 'burn up' on overload, as a poor mans fuse - but that has a rather ill-defined I2t characteristic, and provides an ignition source, and doesn't protect the heater-cathode insulation if the tube is not at fault. Fuses are pretty quality consistent with respect to fusing I-t characteristic, and can be fairly well matched to a PP loadline.
The thermal mass of the fuse element is likely to attenuate the voltage drop parasitic signal that adds to the fixed bias level in a fixed bias PP stage. Given that a hot 2A fuse is circa 0.11V steady state, I would anticipate any thermally induced parasitic signal from resistance change would be lucky to get in to the 10's of mV, and that would only be likely at low frequencies. The parasitic effect is likely to be even less in a cathode biased PP.
Ciao, Tim
Some protection against overvoltage in the OT (due to over-light loading) can be made by installing antiparallel diodes across the output pentodes in PP. The diode-anode goes to ground , diode cathode to pentode anode (get this right or blow your B+). Series-connected 1N4007s x2 or x3 will do.
This is somewhat frequently used in guitar amps. The Laney LC15 is notorious for blowing OTs like this, and adding the diodes fixed a bad example, IME.
The idea is that when the negative-going end of the PP gets to 0V, the diodes conduct. This helps to limit the positive-going end to B+ x2.
It's helpful, cheap, and doesn't get into the way of normal operation. Whether it can protect a fully unloaded PP against overvoltage is doubtful, though.
This is somewhat frequently used in guitar amps. The Laney LC15 is notorious for blowing OTs like this, and adding the diodes fixed a bad example, IME.
The idea is that when the negative-going end of the PP gets to 0V, the diodes conduct. This helps to limit the positive-going end to B+ x2.
It's helpful, cheap, and doesn't get into the way of normal operation. Whether it can protect a fully unloaded PP against overvoltage is doubtful, though.
???
Who suggested doing that?
Nevermind, I see it now. Agreed, not a good idea unless the screens happen to be tied in UL.