basic electronics question: please help me understand this zener

[hg_man]

After several blind alleys with working on the power supply for this *** Behringer POS console, I think I am finally understanding what is happening.  I'm hoping that you all can confirm or deny my logic. 

Here's the schemo.  I'll go through the components one by one to say what I understand they are doing, and you guys can tell me where my assumptions are wrong. ;-)

T, the transformer, is providing ~52V AC.  The "listed" voltage on the PCB is 48V, but my measurement is consistently about 52V.

F, the fuse, is rated at 500 mA.  Theoretically, it protects the rest of the components from excessive current. 

D11, D16, D21, D30 are diodes forming a full wave rectifier.  They make the 52 V AC into ~50 V DC, with ripple.  The reduction in voltage is due to the voltage drops across each pair of diodes as they become the active pair. 

C22 removes the ripple from the 50 V DC, making it 70.5 V.  This 70.5 V is the peak voltage of the 50 V(rms) supply.

C11 - I believe this is the bypass cap for the voltage regulator

LM317HV - this is a high voltage version of the LM317 voltage regulator, with output rating up to 60V.  Its job is to keep Vout at some portion of Vin, based on the development of 1.25V between the Vout terminal and Adj.  The level of Vout therefore depends on the ratio between R2 (240) and R1 (9.1k) (It's unfortunate that the component numbers from the PCB are exactly opposite how I usually look at voltage dividers, which have R1 on the top and R2 on the bottom).  As long as Vin is below 48V, (R1 / (R1+R2))  gives us numbers below 1.25V. When Vin reaches 48V, Adj reaches 1.25V and keeps it there even as Vin continues to rise. 

C1 - bypass cap for the adjustment terminal

D6 - protects the LM317 from discharge of C1

R2, R1 - form the voltage divider that establishes the adjustment voltage between Vout and Adj. 

C15 - bypass cap for the output

D7 - protects LM317 from negative voltages that might end up on the +48V bus

C16 and R3  - filtering for +48 buss output

D1 - this is the tricky bit.  What does it do?  In the application notes from National, they show a similar configuration,  but with a diode in place of the zener.  Its function is listed as protecting the LM317 from the discharge of C1.  How is it suppposed to do that when it doesn't even turn on until Vout is 33 V higher than V in, i.e. 81 VDC!!! 

I've been struggling with this supply for a while now, and it's just bugging me so much that I can't really just let go of it.  I think I may have bought the wrong size of 33V zeners, because when I power up the unit, everything is fine if there's no load at all, but the moment there's any load at all (like connecting just the cable, not the console, the breaker blows up. 

I see some options:

- get a zener rated at 5W

- use a 1N4001 diode in place of the zener

- remove and leave out C15 and its protective diode D1                       


Thanks for your help! 

 

[abbey road d enfer]

The Zener diode does exactly what a standard junction diode would do, protecting the regulator from reverse voltage from the output cap; in addition, it protects the regulator against overvoltage, in case the output is shorted.
IMO, your problem is not related to the Zener, looks more like a short somewhere...

 

[hg_man]

A little background:

- the power supply is in a separate 2-RU box, with a long (5-meter?) cable and connector for connecting it to the console. 

- after several rounds of replacing regulator, fuse, and zener, I pulled all the components (except the transformer) and replaced them. 

Here are my observations:

- When in "test mode" on the bench, with just the transformer plugged in and no cable connections to the power supply PCB, I see 47.2V DC at the output.

- When I attach the cable connections to the PCB and turn on the supply, the fuse blows and the zener also blows - I can see the little "pop" of light in the glass cylinder as it gives up the ghost.  When the zener blows up, it shorts. 

- sometimes when this happens the regulator also fails, as evidenced by being shorted when removed from the PCB

- When I replace the regulator, the zener, and the fuse, I see 47.2V at the output again

- The cable has 7 wires.  When I plug in the ground, I still see 47.2V at the output.

- When I plug in the violet wire to the +48 supply and turn on the box, the zener and fuse blow up, as before.

- I have checked the cable from here to breakfast for shorts. None seen.

- I removed a half-meter of cable from the connector end and resoldered the connector.  This didn't help.

- The zeners I'm using are 33V (as specified by the PCB lettering).  No power rating was specified.  I'm using 1W zeners.  I'm wondering if maybe I need a higher power rating.  The original was in a DO-41 package, just like these, but could have been rated higher I suppose.   

I'd like to just throw the thing in the garbage, but it's become a challenge...

 

[PRR]

> When I plug in the violet wire to the +48 supply and turn on the box, the zener and fuse blow up, as before.

You have a hard-short in this cable.

> None seen

Did you ohm-meter it?

It's not making sense except as a hard-short.

If you need more proof: put a 1K 5W resistor between C22 and the D1 U1 junction. Now it won't blow fuses and is unlikely to blow anything else. If the 1K sits there, cooking, with 70V in and 2V out (317 and D7 drop), you now have time to meter things and confirm (or disprove) that the output is dead-zero, a hard-short.

Your supply is likely fine except it is NOT short-proof. Where did you get the idea to put a 33V Zener there? It does "protect" the '317 from excess in-out voltage differential.... by dumping any excess above 80V input into the Phantom load! And if the Phantom load is shorted, the '317 limits its current, but the Zener will try to pass infinite current into the short.

It only makes sense if the load is absolutely short-proof. Maybe inside a sealed box. Not when feeding a plug, cable, and mystery loads.

 

[hg_man]

You have a hard-short in this cable.

That's what I concluded, but the ohmmeter said otherwise.


Did you ohm-meter it?

I did.  I'll do it again. 

It's not making sense except as a hard-short.

If you need more proof: put a 1K 5W resistor between C22 and the D1 U1 junction. Now it won't blow fuses and is unlikely to blow anything else. If the 1K sits there, cooking, with 70V in and 2V out (317 and D7 drop), you now have time to meter things and confirm (or disprove) that the output is dead-zero, a hard-short.

I'll try that. 

Your supply is likely fine except it is NOT short-proof. Where did you get the idea to put a 33V Zener there?

That wasn't me who designed it that way - that was the nice ""engineers" at Behringer.  There is a diode shown in this position in the application notes.  Maybe one of the engineers needs glasses. 

It does "protect" the '317 from excess in-out voltage differential.... by dumping any excess above 80V input into the Phantom load! And if the Phantom load is shorted, the '317 limits its current, but the Zener will try to pass infinite current into the short.

Sounds like a good way to sell more microphones...



It only makes sense if the load is absolutely short-proof. Maybe inside a sealed box. Not when feeding a plug, cable, and mystery loads.


Ours is not to wonder why, but mine is just to fix the bloody thing.  Either that or use it as a boat anchor.  It might even be worth getting a boat to do that with. 

Thanks for the advice - helpful as always! 

 

[hg_man]

Revelation!  Holy guacamole - it all makes sense now. 

I tried just putting a 1N4002 in place of the zener.  Everything seemed great, and then I found that the chassis (and heat sink) was at +48, until I grounded the chassis, whereupon the +48 supply would go away.  Of course I said, "WTF!", and looked around some more to find that pin 2 of the LM317 is attached to the heat sink - which is now attached to the heat sink, because I must have lost / misplaced / forgotten the little paper barrier that insulates it... Suddenly it all makes sense.