[Amps] HV Diodes

Steven Katz stevek at jmr.com
Wed Oct 8 11:44:10 EDT 2014


I agree with Carl.

Unless a power supply circuit board is very special, the diode leads are a better heatsink than PCB copper foil is.  The only important parameter there is Rthj-a and the wire leads of a power rectifier are better than microscopically thin copper foil on a printed circuit board that isn't thermally conductive.

BTW "potting" can actually improve a lot of stuff if it's done right.  The old Semtech assemblies previously referenced were made from highly thermally conductive potting material from Emerson-Cuming, and the modules that had built-in aluminum mounting rails like the SDHD and SDHD-KM and -KS series had even more thermally conductive potting material, allowing a lot of the junction heat to be transferred to that aluminum rail, which in turn should be tightly bolted to a heat dissipating surface (aluminum chassis, heatsink, etc).

The typical old Semtech rectifiers used in those didn't have Vf = 0.7V/junction at rated current...they were thicker "suprataxial" junctions (not double diffused or conventional epitaxial) and had somewhat higher Vf, typically close to 1.0V/junction at rated current (and at room temp, 25C).  But running them hotter drives the Vf down lower, and while that also drives Ir (leakage) current up higher, those were very low leakage devices that would not run away thermally even at Tj > +175C.

73 WB2WIK

-----Original Message-----
From: Amps [mailto:amps-bounces at contesting.com] On Behalf Of Carl
Sent: Wednesday, October 08, 2014 8:37 AM
To: ka4inm at gmail.com; amps
Subject: Re: [Amps] HV Diodes

I wouldnt use smaller than a string of 1N5408's as they have a much higher surge current capability as the rather obsolete 1N4007 in the era of much higher filter cap C and dirty AC utility lines.

The diode leads are the heat sink and a minimum of 1/2" in free air is recommended by the manufacturers. Ive no idea how potting would modify that. 
My own method is to simply mount them on perf board when possible and thats easy in the Alpha 77DX in place of the "petrified blocks".

Carl
KM1H


----- Original Message -----
From: "KA4INM radio" <ka4inm at gmail.com>
To: "amps" <amps at contesting.com>
Sent: Wednesday, October 08, 2014 9:36 AM
Subject: Re: [Amps] HV Diodes


> On 10/08/14 06:35, Chris Wilson wrote:
>
>   When companies pot a string of 1N4007's (1 kV at 1 Amp) they rate the 
> string at .75 Amp.  The diodes, lacking the posts or copper foil to act as 
> heat sinks, run hotter when potted.  Short leads to the circuit board 
> copper conduct heat away from the silicon die better than long leads.
>
>> Hello Han,
>>
>> On Wednesday, October 8, 2014,  you wrote:
>>
>>> hi George-san and all
>>
>>> I think HV6-1 is equivalent to GI HVPR10-MX12.
>>
>>> I have measured forward voltages of these diode modules
>>> using a CVCC power supply with 100-mA current.
>>
>>> Assuming each of the diode in a string exhibits 0.7 V forward
>>> voltage one can presume how many series diodes are in
>>> the module.
>>
>>> GI HVPR10-MX12 6.36 V: 9 diodes
>>> EDHD 15k 10.67 V per a leg: 15 diodes x 2 (which I am using)
>>> SEMITECH KV-PAC SDHD-15k 12.97 V per a leg: 18 diodes x 2
>>> Toshiba S5166A (13 kV 500 mA) 14.0 V: 20 diodes
>>
>>> I think 10 x 1kV 1A diodes in series is OK as an alternative.
>>
>>> de Han JE1BMJ
>>
>> Has  anyone  experimented  with  home "potting" diode strings to make
>> their  own  blocks?  I  am  curious  about diode lead length, physical
>> orientation, and potting compound used. thanks.
> -- 
>   Ron  KA4INM - Youvan's corollary:
>                 Every action results in unwanted side effects.
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2014.0.4765 / Virus Database: 4037/8348 - Release Date: 10/08/14
> 

_______________________________________________
Amps mailing list
Amps at contesting.com
http://lists.contesting.com/mailman/listinfo/amps


More information about the Amps mailing list