This spark gap came with the transformer. It's a pair of steel spheres about 1cm diameter mounted on the threads of machine screws. These are in turn attached to the terminals in the insulator block of the transformer and it is across the primary. 1cm ball is large but the power level is up to 900W so I guess they did not want to burn the electrodes. haha plus 1947 engineering, no nickel and dime cost-saving just good stuff.
Home made gaps can be made from the round heads of machine screws, or the other (threads) end can be ground to a somewhat hemispherical or smooth shape. A suitable minimum size is a typical rack screw, like a 10-32.
Then set the gap for 1.5-2 times the expected voltage or whatever is less than the transformer's withstanding voltage. So, if you have a 2KV supply and push-pull then a 5-6KV gap is probably OK. It's not an exact science. One thing is the electrodes should not have sharp points or edged but be smooth, or else they will fire too soon. It's always across the winding, never from one side to ground, so the arc can not be sustained by DC but only happen on half-cycles.
What is the formula for arcing? 30KV to the inch? It turns out there are different answers from different situations including things like the diameter of the spheres or shape of the points. In the few I home made I had to experiment. That's why they are adjustable even the professional ones, unless it is a tube and has a stated voltage.
A neon sign xfmr and HV AC meter and variac with a small lamp like 40W between the variac and NST worked ok for that. The lamp limits the power so the gap does not get burned and roughed up. Increase the voltage watching the meter til the arc occurs, then readjust the gap until satisfied. Gaps must be kept clean and smooth.
I think this is a good table here but it starts with 25mm spheres.. who uses that? - from The CRC Handbook of Chemistry and Physics (62nd edition) (left column) and A. Bouwers: Elektrische Höchstspannungen, Springer, Berlin 1939 (right column):
Jochen's High Voltage Page : Measuring high voltages by spark length
Magz: Spark gap is a good idea with the type of HT you are running. The modulation transformer and modulation choke were protected with spark gaps in broadcast transmitters. The gap was normally 3/8 in on a 5 kw transmitter where the plate voltage was around 5 kV and 10-12 kV peaks were expected with positive modulation.
Opcom, the screw head spark gap idea is good. I have used that in transmitters and antenna coupling units. One of those saved an expensive vacuum capacitor in a antenna coupler during a lightning strike. Use brass screws...the 1/4x20 type are commonly available, and a suitable support can be made from sheet copper or brass...
In the audio amp, the thing should never have to "fire", but it should save some expensive iron if unexpected HT transients occur...
Opcom, the screw head spark gap idea is good. I have used that in transmitters and antenna coupling units. One of those saved an expensive vacuum capacitor in a antenna coupler during a lightning strike. Use brass screws...the 1/4x20 type are commonly available, and a suitable support can be made from sheet copper or brass...
In the audio amp, the thing should never have to "fire", but it should save some expensive iron if unexpected HT transients occur...
Thanks, guys. I do have some expensive iron to protect!
Speaking of which, the shipment from Monolith has been in Newark airport for the past week still waiting to go through customs. Any day now...
The panels I had Front Panel Express work on will be delivered back to me today. Once I get the trannies and can double check the fit of everything the panels will then ship back to Landfall to get anodized.
Then the real fun begins.
Speaking of which, the shipment from Monolith has been in Newark airport for the past week still waiting to go through customs. Any day now...
The panels I had Front Panel Express work on will be delivered back to me today. Once I get the trannies and can double check the fit of everything the panels will then ship back to Landfall to get anodized.
Then the real fun begins.
Regarding the spark gap, I was thinking one of these might be cool to have in the amps:
File:Spark-gap-tube.jpg - Wikipedia, the free encyclopedia
There are some for sale on ebay right now, with 4kV breakdown voltage.
Cs137 has decayed considerably by now, though, so may not ionize reliably anymore.
.
File:Spark-gap-tube.jpg - Wikipedia, the free encyclopedia
There are some for sale on ebay right now, with 4kV breakdown voltage.
Cs137 has decayed considerably by now, though, so may not ionize reliably anymore.
.
Last edited:
Looking around on Digikey, I found these gas discharge arrestors which look like they'd work well.
Spark Gaps
They will breakdown at 3600VDC and crowbar the supply, blowing the fuse in short order. They're small, simple to install, and cheap. Seems like they'd be ideal, right?
If I was to install one of these, there are several places it could go:
1) secondary of PT before the rectifier (AC)
2) after the rectifier (AC + DC)
3) at the end of the power supply right before the OPT (DC)
Any suggestions as to the best location? 1) seems to give the most complete protection at first glance. Since 1) is a pure AC application, I assume the arc would extinguish at the next AC zero crossing, whereas it it might continue indefinitely (or at least until fuse blowing and rapid cap discharge) at 3) due to the DC nature of the voltage. 3) sounds like it might cause more damage...
Anyone with experience here care to comment?
Spark Gaps
They will breakdown at 3600VDC and crowbar the supply, blowing the fuse in short order. They're small, simple to install, and cheap. Seems like they'd be ideal, right?
If I was to install one of these, there are several places it could go:
1) secondary of PT before the rectifier (AC)
2) after the rectifier (AC + DC)
3) at the end of the power supply right before the OPT (DC)
Any suggestions as to the best location? 1) seems to give the most complete protection at first glance. Since 1) is a pure AC application, I assume the arc would extinguish at the next AC zero crossing, whereas it it might continue indefinitely (or at least until fuse blowing and rapid cap discharge) at 3) due to the DC nature of the voltage. 3) sounds like it might cause more damage...
Anyone with experience here care to comment?
I have melted several of these devices! We used these kind of devices for lightning protection on remote radio receivers. The paging receivers were mounted on interstate billboard signs to control the lights. They operated from unswitched AC line and were fused with a large enough fuse to keep from catching fire, but wouldn't blow on extreme or extended overvoltage.
CP Claire made the ones we used and it was 1995. These things work well provided the energy being absorbed is low enough to keep the devices below 90C. Once they heat up, they don't shut off leading to total meltdown. CP Claire made a version rated to be wired directly across the 120 volt AC line. They worked great.
They came out with a new version for 240 volts which would make installation of our equipment easier, no neutral connection. We switched to 240 volts. Devices started failing. These devices proved to be incapable of direct connection to a 240 volt line in remote locations and were withdrawn from the market. We switched back to 120 volts.
The real concern is failure of the OPT due to an open load. This is a big issue in P-P amps, and any tube amp that can be driven to cutoff without a load. This causes a huge voltage spike across the OPT leading to arc over. I would look for a device, or series combination of identical devices that can be wired directly across the OPT primary, or possibly the secondary.
I would not put one of these anywhere on the secondary side of the PT, you are just looking for an incident there. IF you can find one rated for DIRECT APPLICATION ACROSS the AC LINE, then you can place one on the primary of the PT if it can survive blowing the fuse. They might exist now, it has been nearly 20 years.
Thanks, George.
I have 5kV, 600mA microwave oven fuses for overcurrent protection, what about putting the fuse in series in the B+ line before the OPT and one of the above devices to B- after it. That will be about 2300VDC in normal operation. If a big voltage spike occurs on the OPT primary and the arrestor fires that ought to protect the OPT and the fuse should blow if the reservoir cap starts to discharge.
Sound reasonable?
I have 5kV, 600mA microwave oven fuses for overcurrent protection, what about putting the fuse in series in the B+ line before the OPT and one of the above devices to B- after it. That will be about 2300VDC in normal operation. If a big voltage spike occurs on the OPT primary and the arrestor fires that ought to protect the OPT and the fuse should blow if the reservoir cap starts to discharge.
Sound reasonable?
I received the front panels back from FPE today - they came out very nice.
I made the holes with hole saws before I decided to send them for engraving, powder coating and infilling. The big hole is for a Simpson DC ammeter and the small one is for a remote low voltage switch for the 833 B+ soft start.
I made the holes with hole saws before I decided to send them for engraving, powder coating and infilling. The big hole is for a Simpson DC ammeter and the small one is for a remote low voltage switch for the 833 B+ soft start.
Attachments
Magz: Agree with Tubelab: Do not put any devices across the secondary of the HT power transformer. Remember there is a 1.41 multiplier to get peak volts involved...Use the 5 kV 0.6 A microwave oven fuse in the B plus line feeding the OPT primary. Use a bleeder resistor across the filter capacitor bank to self discharge it. The LED in series with the bleeder resistor is a good reminder HT is present.
Yes, I have bleeders across each cap. I won't put anything across the PT secondary (thanks for that advice Tubelab and rmb). The filter is choke input so the multiplier is 0.9, correct?
My last proposal above was to put the GDT across the OPT primary, directly after the 5kV fuse, to give both overvoltage and overcurrent protection; this type of setup is mentioned in some of the application notes for GDTs.
Yup, got it. I was concerned with the DC because I gave up on the idea of putting anything on the PT secondary (thanks to you and Tubelab).
So rmb, you agree that an appropriate GDT across the OPT primary preceded by a 5kV fuse in the B+ would be a suitable protection mechanism? It's about all I can fit in there at this point...
That is the accepted rule of thumb. Remember that with loads below the critical current the voltage can be nearly 1.4 X the PT secondary voltage. The bleeders alone will not draw critical current, the idle current through the 833A's should be well above the critical current, so as long as the tubes are hot and the bias is applied before switching on the B+ you are OK.
I have never seen a fuse in the secondary circuit of a microwave oven. I have only looked at the cheap ovens people throw in the trash for some "experiments". A modern microwave oven uses a rather unique circuit where the magnetron is actually the second diode in a voltage doubler as well as the load. The capacitor is way too small to be an effective filter, but forms a resonant circuit with the inductance of the PT secondary. The circuit can deliver a lot of power to the tube in normal operation, but if the tube experiences a high VSWR due to the user operating the oven with nothing inside (or the blonde one nuking a SOS pad to make plasma) the overload reduces the circuit "Q" limiting the current to the tube. The tube runs on half wave pulsed DC. There is a "zero crossing" of sorts to extinguish an arc.
Most fuses will explode if used on pure DC at a voltage high enough to sustain an arc inside the fuse. I know from experience that my 650 VDC power supply will blow a cartridge fuse rated for use on a 600 VAC 20 amp power line in half! An arc will strke as the element inside blows and continue as it eats up the entire element, leaving an arc from one end cap to the other causing a rapid temp and pressure rise inside the fuse leading to something ugly. The arc may continue across the fuse holder after the fuse ceases to exist. Fuses rated for DC are expensive because they contain chemicals or gasses (sulfur hexaflouride) to extinguish an arc.
This puts the device from B+ to ground. It might self destruct before a fuse can blow, or worse.
That would be my choice, but I must state that I have not actually tried it yet.
A large GDT across the secondary may also be a valid option, especially at your B+ voltage levels.
I spent about 5 years with the SE fever building bigger and bigger SE amps up to the 833A breadboard that was never built into a complete amp. I have been slowly working my way down a similar path in the P-P amp world. I have an amp that makes 125 WPC with no protection devices, but it uses $16 OPT's and a $15 PT so I don't care if it blows, therefore it doesn't.
I am working toward two big amps with expensive parts, so it's time. I think I will get some of these devices and experiment. The old ones that I used emitted a purple light when they fired, so it may be possible to sense the light, and possibly the device temp and take the proper limiting action.
Remember this. Theory states that the plate voltage in an SE amp will swing from near zero (the drop across a saturated tube) to nearly 2 X the B+ voltage. The scope reveals this to be true as long as the amp remains in class A (not clipping) AND the load is a pure resistor. This is not the case with a speaker load, especially around woofer cone resonance, and in clipping.
It has been stated that operating a tube amp without a load will fry the OPT. This can be true, and I have indeed set a P-P OPT on fire when my load resistor blew.
The scope also tells me this. If an SE amp remains well below clipping the plate voltage swing will remain within the normal 0 to 2 X B+ limits even with no load at all. Once clipping sets in you are trying to stop the current through an unloaded inductor which WILL cause HUGE voltage spikes. It also takes less drive to push an unloaded amp into clipping.
The situation is a bit more complicated for a P-P amp that doesn't stay in class A. I have seen 2KV spikes on the plates of the output tubes in a guitar amp with 500 volts of B+ in NORMAL OPERATION into a speaker load.
If your amp generates far more power than needed, a simple resistor across the output (50 to 100 ohms) may provide a simple means to avoid OPT meltdown.
If you haven't seen this thread, read it. This guy started down a similar road and even made some sound before his OPT fried. My guess here is that his OPT winder did not really understand what was involved in making an OPT. I have a pair of expensive Plitron OPT's rated for 400 watts at 20 Hz for my "big one." I don't want to blow them.
http://www.diyaudio.com/forums/tube...voltage-driver-ab2-operation-gu81m-tubes.html
BTW....it's too late for my mid life crisis. I decided to build a car instead.....it was never finished.
They need to have fuses in UK models - maybe the regulations vary in other nations.
Fuse position is:
http://blog.earthshod.co.uk/?page_id=155
Fuse and holder usually look like:
MICROWAVE OVEN HIGH VOLTAGE FUSE. COMPLETE WITH HOLDER 5000v 5kv [Micro Fuse 5Kv] - $12.50 : Cellit.com.au, Electrical parts, accessories and components
Even the cheap ones I have dismantled have this protection, so I suspect EU regulations in force!
I built/rebuilt one or two of those in my time, but over time (and with age) I lost my taste for crawling around under cars, skinned knuckles, grease in every pore of my body, so I moved on to electronics. They ranged from a 1970 Pontiac Firebird Formula 400, a 1972 Triumph TR6 and a 1998 Dodge Neon R/T, quite a gamut.
Thanks for all your help, BTW. I'll be interested to hear how your GDT experiments turn out.
.
Last edited:
George, I have no doubt you are correct in your assessment, but for my edification, why not put one of these devices from B+ to ground? All the crowbar circuits I've seen implemented in amps go from HV to ground. The GDTs I linked to are rated for 10,000A discharge for a one time hit and 3,000A for 10 hits. Surely the 600mA fuse would blow before these self-destruct, no?
Why not? Mostly because I'm not sure what will happen when it tries to eat all the energy in your main filter cap at once. 1/2 CV squared. Note that the voltage is squared.
I don't remember how big your cap is but 100uF on 2500 volts is 625 Joules. I scattered banana goo all over a warehouse with a 400 Joule defibrillator (15 uF charged to 5200 volts).
I used the defib for simulated lightning strike testing, and of course "tested" several things including the GDT's.....and a banana. I know that the 120 volt rated GDT's made for phone line use exploded violently with a single hit from the defib, nothing left in the PC board but wire stubs and a black spot. I don't remember all the details since it was nearly 20 years ago, but I do remember needing a ladder to get banana goo off the ceiling.
Why? What it do for you. If you have a reasonable electrolytic and a film cap in parallel, they will keep the voltage at the B+ terminal of the OPT constant. Any surge coming out of the PT will be absorbed by the caps. Any incomming surge big enough to change the voltage in your caps by more than a few volts should blow the line fuse.
I did a 68 Camaro convertible, a 66 Mustang convertible, and the 73 Challenger with an aluminum headed 440, that was never finished. I helped the guy that sold me the Camaro finish his 71 Vette. He would up selling it and buying a new Vette in 85.
It's 333 Joules, 126uF and 2300V.
I see your point about the caps absorbing spikes. With the GDT across the OPT primary, though, where does a spike go, through the output tube to ground. I suppose that can dissipate quite a bit of energy for a short period, but maybe not enough current to blow the fuse on the B+. So the tube ends up being a damping resistor of a sort.
.
I see your point about the caps absorbing spikes. With the GDT across the OPT primary, though, where does a spike go, through the output tube to ground. I suppose that can dissipate quite a bit of energy for a short period, but maybe not enough current to blow the fuse on the B+. So the tube ends up being a damping resistor of a sort.
.
Last edited:
"The situation is a bit more complicated for a P-P amp that doesn't stay in class A. I have seen 2KV spikes on the plates of the output tubes in a guitar amp with 500 volts of B+ in NORMAL OPERATION into a speaker load. "
Can you explain why? I don't get how. Where is the extra 1.5Kv coming from?
But I'm having a hard time catching on to allot of what you explain george (no fault of yours, understand) but I'm tagging along.
Can you explain why? I don't get how. Where is the extra 1.5Kv coming from?
But I'm having a hard time catching on to allot of what you explain george (no fault of yours, understand) but I'm tagging along.