Peavey Pwr Supply Diodes - Why?

[CJ]

any Peavey engineers out there? 

what do these diodes do?

(circled in red)

Thanks!

bonus question- 5 pts - what about the diodes under the voltage regulators?

 

[bruno2000]

I'll go for the bonus, they raise the regs up 0.7V so that the output is 16V (15.7V) rather that 15V like the reg says.
Best,
Bruno2000

 

[Kingston]

bruno got the first one right. Here are the red circles: http://groupdiy.com/index.php?topic=440.msg5340#msg5340

 

[JohnRoberts]

any Peavey engineers out there? 

what do these diodes do?

(circled in red)

Thanks!

bonus question- 5 pts - what about the diodes under the voltage regulators?

They're both part of the same strategy... If you go into the way back machine, there used to be a problem with early 3-terminal regulators latching up if one regulator was reverse biased at start up. The diode in the ground leg and across the supply simply prevented that latch up mode.  This was more of an issue with half wave rectified power supplies, but once bitten some engineers keep putting in these parts, until forced to stop. Having production products that randomly latch up in the field, leaves a lasting impression on mass market design engineers (sometimes even leaves a scar).  :'(

I had to just about threaten an experienced analog engineer in my mixer group to stop adding those diodes because they were no longer needed. I told him to prove to me that they were still needed , he couldn't (but he sure tried) so he finally stopped using them.  The 4 diodes were not that expensive, if they actually did something useful, but if not, its just throwing pennies on the ground and adding more parts that can fail or be assembled wrong. 

His name was Jack... a really brilliant engineer but he never took my word for one single thing...  8) 8) I respected that, I could have used more engineers like him.

JR

 

[Audio1Man]

This would protect the voltage regulators from failure from reverse voltage.  There maybe some other circuity connected that they needed the diodes to protect the voltage regulators.
I would bypass the regulator with a reverse diode. The output regulated voltage may be held up with large storage caps upon turn off and this would back bias the regulators and cause silicon failure.

Duke

 

[JohnRoberts]

This would protect the voltage regulators from failure from reverse voltage.  There maybe some other circuity connected that they needed the diodes to protect the voltage regulators.
I would bypass the regulator with a reverse diode. The output regulated voltage may be held up with large storage caps upon turn off and this would back bias the regulators and cause silicon failure.

Duke

This was a bit of a myth... and probably used to discourage people from adding too much output capacitance. For the output side of a regulator to discharge back through to the input, there needs to be a path to ground for that current to flow into.  A reverse biased rectifier, and reservoir cap is not much of a load on the input side. 

Of course in complex circuits where multiple regulators pull from the same unregulated supply, or it is otherwise loaded to ground or opposite rail,  some reverse current could flow.

JR

 

[Andy Peters]

I would bypass the regulator with a reverse diode. The output regulated voltage may be held up with large storage caps upon turn off and this would back bias the regulators and cause silicon failure.

The ancient 78xx parts had that problem.  The slightly-less-ancient LM317 has, according to the data sheet, a discharge path for that reverse current which is "able to sustain 15 A surge with no problem. This is not true of other types of positive regulators. For output capacitors of 25 μF or less, there is no need to use diodes."

 

[mskeete]

They're both part of the same strategy... If you go into the way back machine, there used to be a problem with early 3-terminal regulators latching up if one regulator was reverse biased at start up. The diode in the ground leg and across the supply simply prevented that latch up mode.  This was more of an issue with half wave rectified power supplies, but once bitten some engineers keep putting in these parts, until forced to stop. Having production products that randomly latch up in the field, leaves a lasting impression on mass market design engineers (sometimes even leaves a scar).  :'(

I had to just about threaten an experienced analog engineer in my mixer group to stop adding those diodes because they were no longer needed. I told him to prove to me that they were still needed , he couldn't (but he sure tried) so he finally stopped using them.  The 4 diodes were not that expensive, if they actually did something useful, but if not, its just throwing pennies on the ground and adding more parts that can fail or be assembled wrong. 

His name was Jack... a really brilliant engineer but he never took my word for one single thing...  8) 8) I respected that, I could have used more engineers like him.

JR

This has happened to me in recent times with 78xx/79 regulators. I switched to LM317/LM337 regulators as I hadnt learnt the solution at the time. Happened to a friend this year or last year and I told him to add the diodes in the ground legs and that cured the problem he was having.

To get a reliable failure at startup I had to leave the equipment off for a few minutes between each test

 

[JohnRoberts]

In hindsight a shorted rectifier diode or shorted reservoir cap, not that uncommon failure modes, could present a low impedance to ground at regulator inputs providing a discharge path backwards through the regulator.

Stuff happens,  I never personally had problems with these, but know people who did.

JR

 

[CJ]

had a NTE959  minus 18 volt reg get shorted to ground with a screwdriver for a split second,

this device is no longer working, short circuit  protection apparently not a strong point with these parts,

better check evilbay for some National Semi old stock,

ST? they are OK i guess, TI? jury still out,

 

[abbey road d enfer]

had a NTE959  minus 18 volt reg get shorted to ground with a screwdriver for a split second,

this device is no longer working, short circuit  protection apparently not a strong point with these parts,

better check evilbay for some National Semi old stock,

ST? they are OK i guess, TI? jury still out,

Long ago, I stopped worrying about the brand of these regulators. They are so well proven there is almost no practical difference, unless they are used at their limits. I'd go for the cheapest available from a reputable distributor (Farnell, Digikey, Mouser...). Along the years I have used Fairchild, NS, TI, ST, JRC, and couldn't detect any significant difference.
The whole series of 78xx/79xx is notorious for the inefficiency of their short-circuit protection, particularly the 24V types.

 

[CJ]

best short protection maybe the old resistor-zener regulator?


maybe less noise too, or not,

here is a blurb on regulator noise vs batts etc>

http://www.tnt-audio.com/clinica/regulators_noise1_e.html

 

[abbey road d enfer]

best short protection maybe the old resistor-zener regulator?

Would need high-current Zeners for this application. Lots of heat.

maybe less noise too, or not,

Zeners are über-noisy, need huge caps.

here is a blurb on regulator noise vs batts etc>

http://www.tnt-audio.com/clinica/regulators_noise1_e.html

This article just confirms common knowledge. What it doesn't cover is the importance of noise currents, that require very thorough design of the "ground" track circulation. Remember that these regulators (particularly 78xx/79xx) rely on brute force for noise control - the output cap discharges all the noise currents into "ground". Noise currents must not be allowed to mix with audio "ground".

 

[JohnRoberts]

had a NTE959  minus 18 volt reg get shorted to ground with a screwdriver for a split second,

this device is no longer working, short circuit  protection apparently not a strong point with these parts,

better check evilbay for some National Semi old stock,

ST? they are OK i guess, TI? jury still out,

Long ago, I stopped worrying about the brand of these regulators. They are so well proven there is almost no practical difference, unless they are used at their limits. I'd go for the cheapest available from a reputable distributor (Farnell, Digikey, Mouser...). Along the years I have used Fairchild, NS, TI, ST, JRC, and couldn't detect any significant difference.
The whole series of 78xx/79xx is notorious for the inefficiency of their short-circuit protection, particularly the 24V types.

Not to quibble but I had a slightly different experience. In a console design where I used 3 terminal regulators on every channel strip. A large console full of around 100 of these (50-7815 and  50- 7915) inside one chassis developed a slight new noise in the noise floor that we eventually traced back to the regulators,  A staticy, infrequent, 1/f (?) noise. After checking regulators from a few other brands, I discovered this was a personal problem with this one brand (Motorola). I looked for a noise spec in their data sheet, and they did not provide one. After a rather frustrating conversation with their field engineers, they declined to admit there was any problem and refused to offer any level noise spec. Ironic perhaps because they were promoting in advertisements about being 6 sigma quality. I guess you can claim low failure rates if you make the thresholds high enough (or use no threshold at all) to define a good part.  :

I ended up black-balling the regulators from that one brand and my noise problem went away.  8)

These regulators were otherwise fine, and the noise in the regulator was probably an engineering trade-off to make a slightly cheaper part.  I do not doubt that my specific design made the regulator noise more of an issue than would show up in typical applications.  The noise floor for a console with over a hundred feeds to the L/R mix should not have regulator noise in it.  ??? Note: This was not significant measurable noise that showed up in noise floor measurements, but was audible to my very picky QA people, doing WFO noise floor listening tests, who wouldn't let me ignore it either.  (We built sound proof rooms out on the factory floor, for them to use when testing these. They kept me honest.)

Just saying there can be subtle differences, while I agree these will be interchangeable for most designs..

JR

 

[abbey road d enfer]

I guess I was just answering CJ's worries about reliability. Indeed, noise issues are a different matter.
I found early that on-board regulators in mixers were a major PITN; indeed they test a console designer's ability to implement a sound ground circulation scheme.
I was bred on on-board requlation; it was deep soul-searching that converted me to using protection resistors. 

 

[JohnRoberts]

I guess I was just answering CJ's worries about reliability. Indeed, noise issues are a different matter.
I found early that on-board regulators in mixers were a major PITN; indeed they test a console designer's ability to implement a sound ground circulation scheme.
I was bred on on-board requlation; it was deep soul-searching that converted me to using protection resistors. 

I like to joke that consoles (and power amps) are the hardest "simple" circuits to design. I've run into funny console power supply behaviors before. But never in the lab when you can easily tweak them... You can't really test them properly without fully loading a frame (36x24).

One early console power supply took off and started singing at 960kHz in sympathy with a nearby AM radio tower (in VA). This later 3 terminal regulator noise didn't show up until well after the console was in full production. Either Motorola tweaked their regulator design, or purchasing changed vendors to save a penny. Either way the buck stopped with me to get the console quiet and the production line back running again.

JR

 

[abbey road d enfer]

I like to joke that consoles (and power amps) are the hardest "simple" circuits to design. I've run into funny console power supply behaviors before. But never in the lab when you can easily tweak them... You can't really test them properly without fully loading a frame (36x24).

I once had a quite heated argument with Doug Self, who had designed a console PSU on the basic design brief of Volts and Amps per rail. Indeed the PSU would fire correctly one time out of 10. A console is a pretty complex and vicious load, like a huge capacitor with a bad case of DA (Dielectric Addiction)

One early console power supply took off and started singing at 960kHz in sympathy with a nearby AM radio tower (in VA). This later 3 terminal regulator noise didn't show up until well after the console was in full production. Either Motorola tweaked their regulator design, or purchasing changed vendors to save a penny. Either way the buck stopped with me to get the console quiet and the production line back running again.

JR

Voltage regulators, by definition have an inductive reactance, just waiting to sing with the capacitor-next-door.

 

[JohnRoberts]

Voltage regulators, by definition have an inductive reactance, just waiting to sing with the capacitor-next-door.

My singing PS problem was more an issue of too much compliance between a clean ground and a dirty power ground. At the moderately high frequency of 960kHz the wire inductance in a 7' wide console was enough to form a resonant circuit...

I fixed it with a handful of .1uF ceramic discs from the local rat shack...  Reality often pisses on theoretical solutions.  ;D

That studio was in the main beam of an AM broadcast tower so every piece of gear in there had problems.

JR