Charge Pump Diode Ratings

[Pavlos]

I've been looking at some of the layouts that require someting other than the usual -9V, +9V or +18V so after a few minutes of hunting for suitable power options I found these which look like they'll be perfect

https://tagboardeffects.blogspot.com/2016/11/charge-pump-extravaganza.html
https://tagboardeffects.blogspot.com/2012/04/boob-tube-v15.html

One thing that does have me slightly concerned and confused though is that the 1n5817 diodes that are mentioned/suggested are only rated at 20V max? Which if they are to be used in a circuit to provide 25V or 33V seems to go against all the usual suggestions to make sure components are rated well above the supply voltage to make sure they don't expire in a spectacular way.

Haven't been able to find anything useful online so far (Mostly just links to these layouts) so a few questions......

First of all, how important is the voltage rating of the diodes in these circuits?
Are the diodes in non voltage critical part of the circuit?
If the voltage rating is important, what effect would substituting have? For example instead of 1N5817 (20V) use 1N5818 (30V) or 1N5819 (40V)

Many thanks in advance

[BetterOffShred]

I got a big bag of 1N5819 for really cheap, and while they say they have slightly higher Fv, none of the ones I've tested seem to have any higher than 5817s..  That being said looking around at all the effects people make, you see 1N4002s and 4007s all over where a 4001 would work... people just built stuff with what they had.   You won't notice a difference in 5819s and 5817s once you factor in a wall wart and peak power grid wall voltage being 116v or something etc..  

As for 1N5817, I use them in charge pumps all the time, but never above 20v.. just get some 5818s or 19s :)

[balazs.bencs]

I agree! I use 1n4004 diodes, cause the local shop here sells them for very little money, once I bought like 500 of them for less than 10 bucks. They work fine for these kinds of tasks in place of the 5817 too.

[Pavlos]

thanks for the quick replies guys

perfect timing for Tayda order, tagged on some extra 1N4001, and added a bunch of 1N4002 and 1n5819 just for good measure.

I've seen mentioned that the 4001's may produce a bit more of a voltage drop compared to the 5817's but I think the possibility of maybe  losing a couple of volts is a sacrifice I'm willing to make, especially if it means I have a safe circuit.

Once the parts arrive, and if all goes well I'll report back with some measurements for the different diodes

[BetterOffShred]

Yeah man, I got a ton of 1N4002 and 5819s that are my go to diodes for power rectification.  You won't have volts of difference, maybe like 50mV ..  

[Pavlos]

sounds good then

For the negligable time and cost to knock up a couple of extra boards, use the same IC and take measurements of each is something that I think might still be worthwhile to do, if only to get a legitimate comparison of the diodes. Of course measuring the caps beforehand would probably help to minimize variables as well, and if nothing else I'll have a couple of spare boards ready and waiting

[Travis]

That would be an even better job for a breadboard, Pavlos. If you don’t have one, I highly recommend getting one

[BetterOffShred]

Yeah I just treated myself to the big bastard 3260 slot one off eBay for 22 clams.  It's so handy.  I wired it up for 3 different voltages and indication plus a power indicator LED .. ooh la la.

Good call Travis

[reddesert]

The 1N400x diodes are all regular silicon power diodes with a voltage drop of about 0.7 V and the difference is in voltage rating: 1N4001 are rated for 50V peak reverse voltage, 35V rms reverse; 1N4002 are rated for 100V peak reverse, 70 rms reverse, and so on. You can get these ratings by reading the datasheet.

The 1N581x series are Schottky diodes with a voltage drop of about 0.3 V and different voltage ratings, 1N5817 is the lowest voltage, then 1N5818 and so on.

The way a charge pump voltage doubler works, the diodes are effectively one-way valves that allow 2 x V_supply to accumulate on a capacitor without discharging back toward the supply. So I think each individual diode is only likely to see 1 x V_supply of reverse voltage. However, it would be prudent to use diodes that are rated for the full output voltage just in case.

What you absolutely must be careful about is using an output supply cap rated for the full supply, with some margin of safety, so for the 33V charge pump circuit I would use an supply cap rated 50V or higher.

[Pavlos]

Hey Travis, thanks for reminding me of my emrarrasingly poor breadboarding skills  something I really should put a bit more effort into I guess, along with following and understanding schematics. Mind you, this could well be a good incentive for me to try and get a bit better at it

 3260 slot breadboard? isn't that a bakery?


Reddesert
Thanks for confirming that, I'd done some trawling around online and with the slightly patchy info I'd found, and my limited knowledge, I'd kinda figured out something like that might be going on (Each stage adds to the previous?) always nice to have it confirmed though I tend to use 35V or 50V caps as standard anyway, they're rarely any more expensive than lower ratings, and often the same size as well so nice safety margin for no compramise.

Could be a good idea to take measurements on all the diodes just to make sure of the voltages they see?

[Pavlos]

Ok, so all the parts arrived and I managed to get it all set up and working on the breadboard, so here are the results.......


I used this layout http://www.geofex.com/circuits/voltpmp1.gif as it was mentioned as the source for https://tagboardeffects.blogspot.com/2016/11/charge-pump-extravaganza.html

Voltages were measured after each diode

Charge pump was a TC1044S from bitsbox https://www.bitsbox.co.uk/index.php?main_page=product_info&cPath=140_166&products_id=3204

All capaciters were 10uF 50v from tayda, I measured a bunch of them and had a quite a few read at 9.3uF so these were used so that the only variables would be the diodes.

9V zinc battery measuring 9.28V

5817
D1 13.91V
D2 18.23V


5819
D1 13.92V
D2 18.22V
D3 22.94V
D4 27.23V
D5 31.97V
D6 36.22V

4001
D1 13.60V
D2 17.78V
D3 22.10V
D4 26.28V
D5 30.60V
D6 34.7F

4002
D1 13.72V
D2 17.74V
D3 22.22V
D4 26.22V
D5 30.64V
D6 34.65V

Looks like the diodes do see an increasing amount of voltage as you progress through the circuit, (Which is why I only did the first stage to 18V with the 5817's) and confirms that the increased voltage drop with the higher rated diodes is negligable, so hardly any compomise to use the 4001 / 4002 and you also have a far better safety margin.

From the measuremens I guess the 5817's are ok for an 18v charge pump, but for anything above that it would probably be a good idea to use the higher rated diodes.

[BetterOffShred]

Badass man.  Yeah the 4002s are great.    Thanks for sharing your findings.

[Pavlos]

With all the brilliant work and effort that others put into the site, the least I can do is give something back whenever my limited abalities allow, so it's a pleasure to be able to contribute something that might be of use to others