KEL84 hot heater rectifier

[ScottF]

I have recently completed a KEL84. I had the transformers and chassis hanging around for a while. I have used a 8A 100v bridge rect for the heaters of the ECF80. Considering the KEL34 used a lower spec'd device I was somewhat alarmed to find that it became too hot to touch within 10 mins (or less). Couldn't measure current but volts were pretty much spot on. I have experimented on the bench with a similar device (GBU8D 200v 8A glass passivated) and despite the 2 ECF80s drawing the correct current at the right voltage the rectifier gets very hot.
I cannot remember noticing this in prev amps (e.g. the KEL34) and find it hard to believe this is normal or at least to be expected.
I would be grateful for any comments and/or suggestions
Thanks, Scott.

[John Caswell]

Hi Scott,
A bit puzzled by this, as generally the rectifier will run warm to hot, but not too hot to touch.
Could I suggest out of curiousity that you knock up a bridge out of say 4 1N5401 type diodes and see if you have the same problem and then report back. It may just be an anomaly of this type/make of bridge,

John aka Dr John

[ScottF]

Thanks John
I had another look at this tonight. The rectifier wasn't too hot to touch but I would have wanted to hold it for too long. Certainly hotter than you'd want a semiconductor to get (thinking V regs and output devices). Using a 10R 20W resistor drew about 800mA (similar to EFC80s) the device was pretty much as before. I then swapped the 4700uF cap with a 10,000uF snap in. I wondered about the effect of ripple current ratings. The rectifier was certainly less hot. I will try this in the amp.
Still puzzled that I am overspec-ing cap & rectifier (orig article used 4700uF cap and 4A 200v rect), but anything for a quiet life and lack of device failure.
I noticed on a post that PCF80s have been used. In this application they would have lower htr current but 2.7 volts down - is this a common approach?
Cheers

[NickG]

I might be wrong, but if the bridge is over spec'd, may it be dropping a higher forward voltage for the current, resulting is higher dissapation?

[Richard]

Check me on this people, about halfway down, "Power dissipation in diodes",
http://www.nhn.ou.edu/~bumm/ELAB/Lec...ture_v1_1.html

I think this is saying that these will run hot by our (human!) standards. If I understand it correctly power is lost in a diode/bridge as heat. The forward voltage loss starts low and increases to a max quite quickly. It's not linear with current but P will equal I x V so if there's a forward voltage loss of 1v then this device (supplying 2 ecf80 at 430mA) will be dissipating .86W and will run hot after a few mins.

Have a look at the "Check valve analogy" to see how the non linear resistance works. A look at a data sheet also shows it, as a curve; at low current draws the voltage loss is less but rapidly reaches a peak.

Better to over-rate btw as the caps can only charge on AC peaks so the rectifier may well be passing 5A or more when charging with a steady draw of .86A.

Rich

[John Caswell]

Hi all,
Well if it is a standard silicon diode (and most bridges are) the voltage drop will be 0.6V at any current up to the max so as Richard says this could be 0.6 x 900mA = .54W X 4 (as there are 4 diodes) = 2.16Watt. So I would expect it to get warm to the touch. But as these bridges are fairly large and black I would not expect them to get too hot.

John aka Dr John

[JerryT]

Originally Posted by ScottF

The rectifier wasn't too hot to touch but I would have wanted to hold it for too long. Certainly hotter than you'd want a semiconductor to get (thinking V regs and output devices). Using a 10R 20W resistor drew about 800mA (similar to EFC80s) the device was pretty much as before. I then swapped the 4700uF cap with a 10,000uF snap in. I wondered about the effect of ripple current ratings. The rectifier was certainly less hot.

Hello Scott

When you initially wrote I was suspecting an additional load on the heater circuit, perhaps caused by some miswiring or a solder bridge as this would generate extra current draw and extra heat.

However 800 mA across 10R implies a DC voltage of 8 volts - rather higher than the specified 6.3 V and outside what I would expect. This contradicts something you wrote earlier about correct heater voltage and current. Assuming this value is right, I wonder if the rectifier is faulty? Or is the transformer supplying a higher voltage than you are expecting? Set your meter to AC and measure the voltage across the input to the rectifier. What value do you get?

Your observations about the capacitor do suggest that one of the problems is the ripple caused by the charge/discharge cycle of the capacitor but I think you need to sort the other problems first.


Jerry

[mickm]

Hi, Happy New Year to you alll...
Re ScotF's post (Num 3 in this thread), I think he maybe refering to me...
I have used Mullard PCF80's in the front end of my KEL84 for three years now and never needed to change one despite 4 to 5 hours average use a day. They are much much cheaper than Mullard ECF80's at radio rallies. I have some industrial Hexfreds - 0.2v forward drop I think - (given to me by a pal) which replace the supplied BR1 diode bridge, and no other change to the cct. I haven't
checked for a long time but I will look to morrow and report the voltage (8v I think under load) and ripple.
It works for me...
Regards,
Mick.

[ScottF]

Dea all

Thanks for your comments
NickG - I think all diodes drop the same voltage so it should be right irrespective of rating.
Richard & Dr John - Yes, the maths does suggest that the BR will get warm, but as I said prev, I can't recall it happening to that extend in the past.
JerryT - ah, well, the devil is in the detail. I said "about" 800mA and was prob being too approximate for this forum
The volts across the 10R resistor were under 7 (but sl above 6.3). It was a diff transformer for bench testing duties just so I could convince myself that the KEL84 tranny wasn't the prob I did check AC (and DC) volts in the KEL84 they seemed OK.
So what happened in the end?
I put in a bigger cap and the BR still got rather hot - I put a heatsink onto it and it's not too bad at all. The amp sounds rather nice too! I might try PCF80s in the future.
Once again - thanks for the support
Scott