Formula Sound FSM-400 repair

[Nakatomi]

Right then.

A long while ago I was given a rather mashed up FSM-400 mixer. I scoped it out with some audio signals & it seems to be working okay but the meter board isn't working right.

Both channels of the meter seem really iffy in that they go between dot & bar mode of their own free will & get dimmer the more level goes into them.

Now, conventional wisdom tells me I can just replace the LM3916 bargraph driver ICs & all will be well. BUT... are these ICs known to just go daft?

I can't seem to find any info on why or how they might've just gone faulty. Power rails look to be OK inside the mixer BTW.

Being that it's a Formula Sound design, I wouldn't expect anything to be operated out of spec so maybe there's been a spike on the power supply side & that took the chips out of whack.

Any experiences with this kind of thing? Anyone happen to have schematics of this mixer lying around? Over to you...

[dicky]

Hi Juski

I'm not familiar with that mixer and don't have a schematic..... however can I ask what is happening on pin 9 of the LM3916 when it is going randomly between bar and dot mode?

If the mixer has a 10 LED VU meter try connecting pin 9 dierctly to pin 3 on the LM3916 if it is not already connected that way - that should force it into bar mode. Dot mode will be pin 9 Open Circuit

The LM3916 outputs are current drains (ie the LEDs are powered from a positive supply and the LM3916 illuminates them by connecting the cathodes to ground. It has it's own built in current limit resistors so you don't need external resistors in series with each LED)

I am guessing the mixer has two LM3916 (10 LED stereo VU meter) and they are both doing the same thing. This, and the fact that the LEDs get dimmer the more that are lit would lead me to suspect you have a PSU problem - so the power supply voltage to the LEDs - and probably the same supply is feeding the LM3916's - is reducing as more current is being drawn. If this is the case it would also explain why the VU meters switch to dot mode - do they do this mostly when lots of LEDs are lit?

Then once they are in dot mode the current drawn is reduced as most of the LDs go off, the supply voltage goes back up, and they jump back top bar mode drawing more current again causing the supply voltage to drop..... and so on. That sounds about the most likely cause of the fault you have described.

Have a look at the voltage supplied to pin 3 and 9 of the LM3916 mate ;) if that is fluctuating I wouldn't be at all suprised if you have a faulty LM78Lxx voltage regulator somewhere on the PCB. If that is the case swap it out for a LM78xx as these are 1A rather than 100mA

Rich

[Nakatomi]

The mixer's own +15 & -15V rails seem fine. The weird thing is that the mode pin, pin 9 is left open & that should, according to the LM3916 datasheet put the chip in dot mode all the time (?!). I've seen working versions of this mixer where the meter is definitely in bar mode. What on Earth can be going on here?

[dicky]

Connect pin 9 to pin 3 mate - see what happens ;)

Also monitor the voltage on pin 3 and the voltage on the anodes of all the LEDs (also check with a continuity meter if pin 3 of the LM3916 and the LED anodes are all connected to each other - most likely they will be

I suspect you have stable +15v and -15v rails to the mixer audio and a separate (probably 12V) rail to the LEDs on the VU meter that also powers the LM3915s and that 12V rail comes via a LM78L12 voltage regulator tapped from the +15V rail.... and that the voltage regulator is faulty


Check the voltage on pin 3 of the LM3916 chips and see if it is varying. If that is not the case then it is hard to daignose this fault via the forum - but if you want to send it to me I will certainly have a look for you :)

Bear in mind the LEDs are not powered by the LM3916 - it is just a 10x LED current drain - so I can't for the life of me see how it could dim the LEDs the more they are lit if the V+ supply to the LEDs is stable... but there again I suppose anything is possible once we rule out the likely then the remainder is going to be the answer.

Rich

[dicky]

And come to think of it.... if the anode voltage to the LEDs is stable - monitor the cathode voltage for the first LED (pin 1 of the LM3916) if that voltage increases the more LEDs are lit then check that pin 2 4 and 8 are properly connected to 0V (they should be)

If that is OK then check the resistor connected between LM3916 pin 7 and 0V is OK and reads the correct value - as that sets the LED brightness

If all these things check OK then change the LM3916 and see what happens

Rich

[Nakatomi]

Connect pin 9 to pin 3 mate - see what happens ;)

Also monitor the voltage on pin 3 and the voltage on the anodes of all the LEDs (also check with a continuity meter if pin 3 of the LM3916 and the LED anodes are all connected to each other - most likely they will be

I suspect you have stable +15v and -15v rails to the mixer audio and a separate (probably 12V) rail to the LEDs on the VU meter that also powers the LM3915s and that 12V rail comes via a LM78L12 voltage regulator tapped from the +15V rail.... and that the voltage regulator is faulty


Check the voltage on pin 3 of the LM3916 chips and see if it is varying. If that is not the case then it is hard to daignose this fault via the forum - but if you want to send it to me I will certainly have a look for you :)

Bear in mind the LEDs are not powered by the LM3916 - it is just a 10x LED current drain - so I can't for the life of me see how it could dim the LEDs the more they are lit if the V+ supply to the LEDs is stable... but there again I suppose anything is possible once we rule out the likely then the remainder is going to be the answer.

Rich

There are +15 and -15V coming in on the meter board connector, the +15 side goes to LM3916 Vcc (pin 3) and the -15V goes through a 6.2V zener to pin 2 - the 0v side of the LM chips (no resistor either), and pin 4. The zener voltage is fine. Volts across pin 3 & pin 2 are about 23.8V, nice & steady. That seems about right to me.

I do think the LM chips are duff, I'm just wary of changing em out incase whatever made them go faulty in the first place is still an issue. LM3916 chips aren't as easy to come by cheaply as they once were. Six quid for a couple off Ebay.. can't afford to keep popping them ;)

It wouldn't be the first formula sound mixer I've ran into with meter issues. Pretty sure I've worked with a PM80 with a similar kind of problem in the past. Maybe these bargraph ICs are just powered a little bit too close to their maximum & even just a slightly bad PSU wibble can send them bad.

[dicky]

Hi Juski

Well as I said once you eliminate everything else then the chip is the next likely. Did you just check the resistor that sets the LED brightness. Also the datasheet says Vled should be below 7V otherwise you need external dropping resistor to limit the power dissipated in the IC - which suggests the LEDs are not powered from the 15V rail directly - so you should investigate that area of the circuit to see what is regulating the supply to the LED common anodes

Are the LM3916 ICs socketted?

I have some LM3915 here (same chip pin-for-pin but the 3915 is 3db steps, the 16 is VU scale with -20,-10,-7,-5,-3,-2,-1,0,+1,+2)

These would prove the whether the fault is the LM3916 before you bought some chips - I'll happily post you one for test purposes if you like.

LM3916 two for £1.40 inc postage if you order from china and don't mind waiting acouple weeks http://www.ebay.co.uk/itm/2PCS-LM3916N-1-IC-DRIVER-DOT-BAR-DISPLAY-18-DIP-NEW-GOODQUALITY-/300812816308?pt=LH_DefaultDomain_0&hash=item4609d74fb4

Rich

[Nakatomi]

hi Rich.

Thanks for the pointers - I really think maybe the power supply had a blip & took the chips past a point of no return. The PSU rails are regulated but aren't clamped so it's a strong possibility.

Everything checks out, so the chips are likely duff. They're not socketed - none of the ICs on the meter board are because of a lack of clearance. I'll investigate whether I can use some low profile sockets though.

Not having decent desoldering gear available I'm going to wait until I start my new job next month. I checked everything else out months ago & the mixer output is more than capable of shoving out a very healthy line output from all mixer inputs. I just couldn't get the thing to clip :D Well not reasonably anyway.. 2V pk-pk into a mic input would do it.

Bar the meter, some new knobs & a new fascia it'll be a very nice mixer. I would've said little mixer, but it ain't ever gonna be that.

[dicky]

Good luck mate

As you need to wait to do the desoldering then may as well order 2 chips from China for £1.40 - they will arrive in the meantime :)

Desoldering by the way - what technique do you prefer?

If it's a multipin device then I generally like to snip all the legs off the device as near to the casing as I can, then pull them out of the PCB individually using long nosed pliers and my soldering iron. Finally I reheat the solder hole and sharply tap the PCB against the anti-static mat and the remaining solder comes out leaving a clean hole.

I find it much kinder to the PCB that way, especially ones with PTH, and once I replaced the device the old one is going to find the dustbin anyway so whatever I to that is irrelivant:p

Rich

[Nakatomi]

Good luck mate

As you need to wait to do the desoldering then may as well order 2 chips from China for £1.40 - they will arrive in the meantime :)

Desoldering by the way - what technique do you prefer?

If it's a multipin device then I generally like to snip all the legs off the device as near to the casing as I can, then pull them out of the PCB individually using long nosed pliers and my soldering iron. Finally I reheat the solder hole and sharply tap the PCB against the anti-static mat and the remaining solder comes out leaving a clean hole.

I find it much kinder to the PCB that way, especially ones with PTH, and once I replaced the device the old one is going to find the dustbin anyway so whatever I to that is irrelivant:p

Rich

Hi Rich. I've already got the chips, but I'll probably order some Chinese ones to be on the safe side too.

As for desoldering through hole - exactly as you describe, except I tend to clean the holes with solder wick or a desoldering pump. If it's a heavy ground plane though I generally pre-heat the board with a heatgun before clearing the ground PTHs :)

SMDs.. I just get right in with the iron, flood it with solder & lift off the IC if it's a SOIC. Otherwise, heatgun if I want the chip back, or just cut it orf with a sharp scalpel. I keep meaning to buy some KwikChip or is it ChipQuik? Anyway, it's an amazing new desoldering alloy you ought to look at.

[dicky]

I will have a look into that :)

I've not actually got a heatgun or had use for one yet but it does sound like a useful tool the way you describe it.

This looks like an interesting tool and not stupidly expensive either (the translated description is a bit of a laugh though) :muppet:

http://www.ebay.co.uk/itm/NEW-T862-INFRARED-REWORK-STATION-SOLDERING-DESOLDERING-STATION-SMD-/281680896545?pt=LH_DefaultDomain_3&hash=item41957d6221

[Nakatomi]

I will have a look into that :)

I've not actually got a heatgun or had use for one yet but it does sound like a useful tool the way you describe it.

This looks like an interesting tool and not stupidly expensive either (the translated description is a bit of a laugh though) :muppet:

http://www.ebay.co.uk/itm/NEW-T862-INFRARED-REWORK-STATION-SOLDERING-DESOLDERING-STATION-SMD-/281680896545?pt=LH_DefaultDomain_3&hash=item41957d6221

:lol:

I tend to use an ordinary hot air blower, with a bit of bent tin on the end to concentrate the heat flow, held on with a jubilee clip.

Nice to see other people are interested in fixing kit around here. You don't run into much repairing of kit out here in internet land these days, let alone modding things to perform better or differently :)

[Nakatomi]

Update!

I took the meter board to work today and found it's the leds at fault. Not all of them but I'm going to swap them all anyway.

My attempt at desoldering a bar graph chip at home without proper tools meant I needed to do some track repairs but they're all good now.

Once the leds are all changed it should all be fine. Silly series configuration to save current. Bah!

[Nakatomi]

Final update....

It's all fixed up now. At last!!

Turned out to be more than just a couple of LEDs needed replacing, but that's what the original fault would've been. In my ham-fisted probing around I'd knackered the TL074 op-amp on the meter board. Oops.

Might buy a new front panel for it & sell it on. I could almost buy a pretty good new one for what these go for 2nd hand.