Behringer ADA-8000 owners: PLEASE READ!!!

this is a thread which has branched off from the topic of Behringer ADA-8000's with dead or dying power transformers.

On investigation of the two which I have here, I suspect that there appears to be a potentially serious design flaw in the power supply, and I think I know how to fix it. It's a $15 fix, though it may void your warranty, so if your unit is still under warranty, -hold off. Once it's out of warranty, you may want to perform the mod... if it turns out that I'm on the right track, at least. :wink:

I'm still looking into the values and details, but the nature of the problem is this: The power transformer is a 230V primary with a 120V tap, NOT a center-tap, and NOT a split-wound primary: It's a 0-120-230V primary.

PART 1: In 230V operation, the entire primary shares the load, and the increased inductance halves the current. In 120V operation, almost half of the primary coil is unused, and the remaining half has to pass DOUBLE the current in order to transfer the same amount of current to the secondary. This means that -all things being equal- 120Volt operation is slightly tougher on the primary.

Fortunately, the primary appears to be rated sufficiently highly to do the job... so long as nothing goes wrong to increase the secondary load dramatically.

PART 2: The secondary windings are just too freakin high. -With a measured 121V RMS coming out of the wall:

The secondary AC voltage feeding the (7805) 5 volt regulators reads 11.2VRMS under load... That means that they are dissipating a LOT of heat... and there are three of them in parallel... right next to the main smoothing electrolytic cap.

The secondary AC voltage feeding the (7815/7915) 15 volt regulators reads 20VRMS under load per leg. The raw voltage going into each regulator is 27-point-something volts DC. They are dissipating a LOT of heat: certainly the heatsink is rather too hot to be comfortably touched.

The secondary AC voltage feeding the 48 volt regulators is 62VRMS. This is rectified up to about 80 volts.

Result: This is fine until one of the regulators decides that enough heat is quite enough, and fails. If it fails hard-short, it loads the rail down, pulling the primary current so high that the primary -which was still managing to cope at 120V- crosses the line into too much, and I suspect that the core saturates. Then the primary just gets hotter and hotter, and eventually it's night-night time.

230V users: If I'm on the right track, you won't have to worry so much about the primary, but your regulators may still be getting just as righteously hot, and they masy still be prone to failure. Even before they fail, they may be 'baking' the smoothing caps.

Can anyone with a 'non-failed' ADA-8000 check this out for me? Do the regulators naturally run too hot to touch?

-If my suspicions are true, I suspect that there may be an easy fix; I've figured out an inexpensive plan which would cool the regulators and also provide improved performance and reliability, -but first I need to check that the hot regulators are normal, and that I'm not instead looking at two units which happen to have some other failure which is heating the regulators excessively.

If I'm on the right track, this is VERY exciting.... -well, for me, at least! :wink:

Thanks in advance...

Thanks for sharing :thumb:

Haven't received the faulty one yet (Europe), but I'll have a look at my 'good' ADA. Will do that ASAP, but that won't be before this Sunday.

Bye,

Peter

Cool.

-And to show my immense appreciation, here's some gratuitous Serge:



Anyone else who has one like to double-check?

Keef

[quote author="SSLtech"]The secondary AC voltage feeding the (7805) 5 volt regulators reads 11.2VRMS under load... That means that they are dissipating a LOT of heat... and there are three of them in parallel... right next to the main smoothing electrolytic cap.[/quote]

In my 230V unit they see 12.96V at their inputs! And run super hot of course.

The 15V regulators see 23.36V and the two with the heatsinks on get quite hot also.

I hope I have a non-broken unit since it has only been used fro about 6 hours since I purchased it about 3 years ago and never showed any problems.

Olaf

:thumb: :green: :wink: :grin: etc etc !

Hi Serge, how are you doing ?

Yes, talking about people no longer with us, that's also the reason I can't have a look more early. No worries, it'll become OK again though.


Talking about S.G. again, this week the 1972 'sister-album' to Melody Nelson
arrived here. It's by Jean-Claude Vannier, the arranger behind the M.N.-album.

A bit weird (intentional), and not exactly like M.N. but highly recommended. One of those rare albums you can play five times a day (that's like it is here now)...

http://www.moviegrooves.com/shop/lenfantassassindesmouches.htm



Ahh, let's forget about those German/Chinese boxes, let's play French music :wink:

Cheers,

Peter

Olafmatt:

-Sounds like I'm onto something. Let me test a modification on one I have here, then I'll be posting it for all to enjoy!

Keith

Could you just imagine Bush with 230v at his disposal!!!!

Ok. Have the ADA8000 since a few years (3 or so i think).

The first unit had some times (every few hours) digital click ... got it exchanged by behringer to a brand new unit incl. cables and other stuff (did send in just the def. unit in non original package). Got the new unit back after 5 days! This was a very positive surprise i didnt expect (its the only Behringer gear i have).

The first unit and also the second unit are on for long times, sometimes a few days without switching off. Its a 230V unit.

I plan to rearrange my rack in the near future, when i do this i check out the voltages. If you need the info urgent ill try to do it before (its pita to crawl behind the rack)

I have repaired a few Behringer units (not the model under discussion though) - with blown mains transformers - but always replace them with a quality unit from RS or Farnell.

As I am in the UK (230 volts) - the whole of the primary winding is used - next time I get one I will have a close look at the tapped winding. Must save Behringer a few pence on each transformer.....

I am always amused at the black plastic tape wrapped around the mains IEC inlet - make a change from hot glue stuff gobbed everywhere!

Behringer - as yes - the 'No schematic' company - surely they can't expect Pro's to use equipment that can't be repaired locally?

Well, I'm testing the modification on the +5V, the +18V and the -18V rails right now.

I still have to calculate the maximum current load for the +48V rail... gimme some time to get things improved.

Basically here's a quick rundown of the mod so far:

undo all the screws on the 8 XLRs and the 2 fiber optic connectors on the rear panel. (18 screws total)
Remove the screws holding the PCB to the chassis (4 screws)
'shuffle' the circuit board up to gain access to the solder side.

Desolder these diodes at the POSITIVE end: (the end with the white band!)
D11 & D10 (positive node of the +5V bridge)

D7 & D6(positive node of the +48V bridge).

D20 & D21 (positive node of the ±15V bridge)

Desolder these diodes at the NEGATIVE end: (the end WITHOUT the white band!)
D22 & D23 (negative node of the ±15V bridge)

Okay, Double check that you've lifted the correct end of each diode. (-On my PCB, the screen printed legend for each and every diode happens to correspond with the end that you have to lift; yours may be the same, but DOUBLE-CHECK!!!)

Bend the free ends of D6 & D7 together and solder them to form a node.

Bend the free ends of D10 & D11 together and solder them to form a node.

Bend the free ends of D20 & D21 together and solder them to form a node.

Bend the free ends of D22 & D23 together and solder them to form a node.

Cut a dozen or so 4" or 5" (100-125mm) lengths of insulated hookup wire, of gauge strong enough to handle an amp or two, (specially for the 5 volt rail!).

Cleanly solder one end of a piece of wire into the now-empty hole which used to be occupied by D6.

Solder one end of a second piece of wire into the now-empty hole which used to be occupied by D10.

Solder one end of a third piece of wire into the now-empty hole which used to be occupied by D20.

Solder one end of a fourth piece of wire into the now-empty hole which used to be occupied by D22.

Re-mount the board in the chassis and replace the screws. Be particularly careful with the screws into the optical connectors: DO NOT force the screw in without supporting the back of the connector strongly, or you can break the pins away from the board... (ask me how I know! ...although not with one of these particular units...)

Now, each rail mod is essentially a reservoir (smoothing) cap followed by a power resistor. The rectifier output of each rail will go across a smoothing cap (the other end of which is referred to 0Volts), then through a power resistor, and back to the board.

The idea is that some ripple current is relieved from the on-board smoothing cap by the newer (and bigger!) capacitor, then the raw voltage will be reduced off-board to a more manageable voltage, moving much of hte heat load from the regulators. In addition, the on-board capacitor now forms the shunt leg of an RC stage, smoothing the voltage still further.

So, the resistor and capacitor values which seem to work so far, are as follows:

+5V:
4700µF/16V
10Ω, 10Watt (ceramic)

+18V:
1,000µF/35V
25Ω, 25Watt (metal body)

-18V:
1,000µF/35V
25Ω, 25Watt (metal body)

So here's how I've done it so far:
In addition to the wires soldered to the lower side before we put eh board back in, we need to solder one more wire, onto the NEGATIVE end (the end without the white band!) of D8. -This is our modification 0V for all of the rails, and we'll be 'star-connecting' one end of the capacitor for each rail modification to it.


+5Volt Rail Mod: securely Mount a 4700µF/16V capacitor inside the case.
Run the wire from 0Volts (the negative end of D8) to the NEGATIVE terminal.
Solder a new length of wire from the junction node of D10 & D11 to the positive terminal.
Solder a second connection to the positive terminal, to one end of the 10Ω/10W resistor.
Locate the unconnected end of the wire going to the hole vacated by D10, and solder it to the remaining terminal of the 10Ω resistor.


+18Volt Rail Mod: securely Mount a 1000µF/35V capacitor inside the case.
Run the wire from 0Volts (the negative terminal of the cap in the +5Volt rail mod) to the NEGATIVE terminal of the 1000µF cap.
Solder a new length of wire from the junction node of D20 & D21 to the positive terminal.
Solder a second connection to the positive terminal, to one end of the 25Ω/25W resistor.
Locate the unconnected end of the wire going to the hole vacated by D20, and solder it to the remaining terminal of the 25Ω resistor.


-18Volt Rail Mod: securely Mount a 1000µF/35V capacitor inside the case.
Run the wire from 0Volts (the negative terminal of the cap in the +5Volt rail mod) to the POSITIVE terminal of the 1000µF cap.
Solder a new length of wire from the junction node of D22 & D23 to the negative terminal.
Solder a second connection to the negative terminal, to one end of the 25Ω/25W resistor.
Locate the unconnected end of the wire going to the hole vacated by D22, and solder it to the remaining terminal of the 25Ω resistor.


So far, it's looking good, but the real test is how long it stays on versus how warm it gets. -Fingers crossed!!!

As soon as I've got the values for the 48V rail calculated, I'll post it.

there's a further protection to this modification: if a regulator ever DOES fail hard-short, the current is now limited by the power resistor which we're installing. This should be enough to protect the transformer... I think we've established that the primary fuse simply isn't doing much to protect the transformer primary!

Keith

(Edited for typing errors, and clarity in the mod description.)

Alright...one issue down...now lets make the preamps sound good!

Keith its sounds like a proper transformer would be the better answer for the USA. I did read AudioJohn's post about using a better transformer

The above mod make sense for 230VAC.

You can also wire a 12.6 etc transformer as a buck to lower the voltage to the primary.

GEOFEX has a writeup for this in a section to drop the current line voltage into older amps to one more like they were designed for IIRC.

Maybe a cheap 115 to 230 converter like the one Klaus wrote about in a thread at his forum for the USA.
A little transformer loss and use more of the transformer wired for 230.

well, as long as the unit's not in unlimited-current failure mode, the primary is adequate at 120V. It's when the regulator (7915 anyone?) fails hard-short that the primary gets hammered. The 230V just gets more protection, that's all.

I reckon that this mod should be adequate if done to US units as a preventative measure. The better alternative would be a transformer with 120V primary, and 8V (for the 5V rail) 16-0-16V (for the ±15V rails) and a 45V (for the 48Vrail) secondaries.

If anyone knows of a 30VA toroidal tranny with 8v, 45V and 16-0-16V, I'm all ears!

Actually I have one transformer with almost those exact secondary windings on it right here... but it's about 150VA... I'll try it out!!!

I'll get back to the 48V rail a little later.

Keith

Okay...

48V rail values:

Capacitor: 150µF/100V
Resistor: 200Ω, 5Watts.


+48Volt Rail Mod: securely Mount a 150µF/100V capacitor inside the case.
Run the wire from 0Volts (the negative terminal of the cap in the +5Volt rail mod) to the NEGATIVE terminal of the 150µF cap.
Solder a new length of wire from the junction node of D6 & D7 to the positive terminal.
Solder a second connection to the positive terminal, to one end of the 200Ω/5W resistor.
Locate the unconnected end of the wire going to the hole vacated by D6, and solder it to the remaining terminal of the 200Ω resistor.

That appears to be it. The load on the 48V rail varies dramatically, from almost nothing (with phantom switched off on the front panel) to 140mA (with phantom engaged and all 8 inputs shorted to ground to simulate the maximum that can possibly be drawn). The 200Ω resistor is the highest that can be used (the raw INPUT to the regulator sags to 48V in this condition) It seems to be the best compromise, since 8 mics all plugged in and drawing phantom still doesn't get the regulator too hot.

Keith

is this a toroid transformer?

I had some old behringer stuff with toroids that overheated and blew their thermal fuses.

I just bought an AD-8000 from Norman_Nomad but I haven't opened it up yet. I'll let you know when I do.

Hey Svart,

If you find a replacement with 6V, 16-0-16V and 45V, by all means replace it, failing that, I'd be really interested in your take in terms of whether or not you agree that the sheer raw power dissipation load placed on the regulators is excessive.

If so, and you can't find a replacement tranny, Maybe you'd like to try the mods, and report back as a 'guinea-pig'?

-If you find a better transformer (lower secondary voltages) though, it'd be great to post a part number. :thumb:

-If you want me to send you up a "kit" of some power resistors and caps so you can guinea-pig the mod, let me know.

Keith

Imagine how there units feel in Australia... Our mains is rated at 240V and floats well above that at times, even in Sydney.

I have a 110V unit here we brought back from the US, I was going to convert it to 220V but no idea where to get an appropriate torroidal, so like all Behringer stuff new or old, it is ultimately landfill.

Alot of the Behringer gear can be switched from 120v to 240v by reversing the fuse holder.

Well, it's been running for about 4 hours now, regulators running as cool as a cucumber. I upped the 5V rail resistor value slightly to 12Ω, and 15Watts. Still enough regulation margin at the 7805 inputs. (almost 8Volts)

Communicating to and from the Digi 002 very happily.

Imagine how there units feel in Australia... Our mains is rated at 240V and floats well above that at times

Click to expand...

The power in the UK is also called "240Volts", but the official typical socket measurement that the power company aims for is 247Volts. The current Federal European standard means that all equipment must accept a ±10% voltage swing, so a piece of gear rated nominally at 230V will accept up to 254 Volts.

Keef

I made a small box with cheap standard transformers (2x18V, 6V). I don't need phantom power, but a small 2x30V will also fit inside this box.
I haven't found a transformer that fit inside the Behringer case, so I went this way and it's rather easy. The box powers two Behringer.


Sorry for the bad quality, cellphone camera.