Debug Holy Roller/Baja LA2A Optical Limiter

[dthurstan]

Hi

I'm hoping someone can give me some pointers with this please. Here's the schematic;


Built a vero from this verified layout;


I've used TL072s for the ICs and 2n5458 for the jfet. Once I built it I found that Vref is going high 7.7v. The voltage at 47k/47k divider (pin 5 IC1) is correct 4.4v. But the output of the buffer goes high?
Here are my voltages:
voltage in: 9.01
D4 cathode: 8.76

Q1 - 2N5458
G. 7.01
S. 8.7
D. 8.76

IC1 - TL072
1. 7.68
2. 7.68
3. 7.68
4. 0
5. 4.38
6. 7.69
7. 7.69
8. 8.76

D2 anode: 7.71
D2/D3: 7.09
D3 cathode: 6.47
LED cathode: 8.1

IC2 - TL072
1. 7.70
2. 7.70
3. 7.70
4. 0
5. 7.69
6. 7.6
7. 8.1
8. 8.76

Here are pics of the component & trace sides.



Thanks
Dave

[duck_arse]

you need to trim those long component legs, and I'd get a scalpel/xacto into all your track-cut corners to clean out the little semi-circles left behind, which may be hiding some unwanted shorts. also, I'd pull the 10uF [C8?] as we are not supposed to load opamp outputs w/ whacking big caps like that.

[anotherjim]

I'd also pull C8. A TL072 is highly strung and may not like a capacitive load. The Vref is already smoothed by C7 so the opamp only has to buffer that voltage.
I think a TL062 is maybe a bit more like the original LF353 if you have one to try.

[dthurstan]

Hi duck_arse, thanks. I left the LED/LDR and the jfet long because I might swap them out if I needed too. You need to play with the LED/LDR combo to get it sounding right. The track cuts should be fine, I tested them all. Will check again though.

Hi anotherjim, thanks. I wasn't sure if it was the IC, I thought there must be something pulling it up but didn't know what.

I've removed C8, still the same. I have a TL062 somewhere I'll try that.

I did do a search to see if TL072 would be ok instead of the LF353, but it depends what you need it for really.

Thanks
Dave

[anotherjim]

No, if C8 wasn't a problem a different opamp won't change anything either.
Maybe something wrong with the connection from output to the -input. That link between pins 6 & 7 - is it actually making contact with the solder side? That Vb buffer circuit should copy the voltage on pin5 to pin7 but only if 6 & 7 connect together.

[dthurstan]

No, if C8 wasn't a problem a different opamp won't change anything either.
Maybe something wrong with the connection from output to the -input. That link between pins 6 & 7 - is it actually making contact with the solder side? That Vb buffer circuit should copy the voltage on pin5 to pin7 but only if 6 & 7 connect together.

Thanks Jim. Already did it last night and it worked!

Very subtle effect, it is a limiter after all. Gonna have a play with LDR/LED & also try & get a little more signal into the 'opto element'.

Dave

[anotherjim]

What did you do last then? Did you fit a TL062?

[dthurstan]

 Yeah, I removed C8 & fitted a TL062. I haven't changed anything else yet.
Btw, what spec/characteristics should I be looking at in an opamp?
I had a look over TL072, TL062 & LF353 sheets, but I didn't know what I was looking for.

Dave

[anotherjim]

Probably bandwidth and slew rate. The 062 is low power and that's usually done at the cost of speed. It can be a good choice where speed or noise isn't so important. Ideally, both halves of a chip like the 062 wouldn't be used in the audio path. The other half could have been used for driving the LED for the LDR, but that might of made the board layout harder to work out.

Is it not possible the 072 you had there was faulty?

I would recommend fitting sockets for your IC's. Unless it's a known good and well made etched pcb, there's always a chance you have to pull a chip out. It can be better to buy turned pin header socket strips and cut them to length. They leave everything clear under the chip for links (or even components) and you don't have to stock different sizes as you would with the normal IC sockets.

[dthurstan]

Hi thanks Jim. Apologies for the late reply, work is crazy at the moment.

Well, I tested the TL072 and it looks faulty on that side.

Yeah, I used to put sockets in, as I thought chips would be pretty sturdily, although this the second faulty one I've found (different chips). Maybe I should handle chips a little more carefully.

Here's a quick comparison of specs;


 
     
 
     
     
     

-	LF353	TL062	TL072
bandwidth	3	1	4
slew	13	3.5	16
CMRR	100	86	86
Icc	3.6	0.2	1.4

[anotherjim]

Ok, from the specs it looks like the 072 ought to be better.

Does seem more likely the 072 was just faulty for whatever reason.

It will still be better off without C8!

[bowanderror]

Recently built the same vero layout w/LF353s and the LED/LDR may take some tweaking to get right. I'd definitely suggest butting the LED right up to the face of the LDR & sealing it with a few layers of black heatshrink tubing. The increased proximity will increase the amount of "squish" you get from it, as well as helping isolate it from stray light. That isolation is super important as any stray light will end up causing gain reduction even when you're not playing.

If you really want to improve the squash, you can:

• Change R19 - Lower resistor value = more LED drive current & more signal shunted to ground (more gain reduction). Higher resistor value = less LED drive current & less signal shunted to ground (less gain reduction)
• Change R5 - This is R1 in the gain-reduction voltage divider, the LDR is R2, so the amount of gain reduction is equal to R2/(R1+R2). You probably want your R6 to be in a similar range what your LDR is when you're playing. Somewhere between 1/2 & 2x the value of the LDR worked for me for R5
• Increase R6 - With the Comp knob up, if you are getting more volume drop than you can make up with the Level knob, increase R6 to 6.8k or 10k to increase the amount of makeup gain. I had to use a 10k in mine.

I found that with my knock-off eBay vactrols, the on resistance was SUPER low (less than 1K), so I had to both increase R19 & decrease R5 in order to use the right part of the LDR resistance vs LED current graph. With homemade LED/LDRs, the rise/fall time of the LDR is also a factor, but testing that is quite a pain.

[dthurstan]

Recently built the same vero layout w/LF353s and the LED/LDR may take some tweaking to get right. I'd definitely suggest butting the LED right up to the face of the LDR & sealing it with a few layers of black heatshrink tubing. The increased proximity will increase the amount of "squish" you get from it, as well as helping isolate it from stray light. That isolation is super important as any stray light will end up causing gain reduction even when you're not playing.

If you really want to improve the squash, you can:

• Change R19 - Lower resistor value = more LED drive current & more signal shunted to ground (more gain reduction). Higher resistor value = less LED drive current & less signal shunted to ground (less gain reduction)
• Change R5 - This is R1 in the gain-reduction voltage divider, the LDR is R2, so the amount of gain reduction is equal to R2/(R1+R2). You probably want your R6 to be in a similar range what your LDR is when you're playing. Somewhere between 1/2 & 2x the value of the LDR worked for me for R5
• Increase R6 - With the Comp knob up, if you are getting more volume drop than you can make up with the Level knob, increase R6 to 6.8k or 10k to increase the amount of makeup gain. I had to use a 10k in mine.

I found that with my knock-off eBay vactrols, the on resistance was SUPER low (less than 1K), so I had to both increase R19 & decrease R5 in order to use the right part of the LDR resistance vs LED current graph. With homemade LED/LDRs, the rise/fall time of the LDR is also a factor, but testing that is quite a pain.

Hey, thanks for this only seen it now.
Will have a play with these values. I need to properly cover the led/let using elec tape at the moment.
I replaced the LDR with a GL5506 which has lower light resistance. I think it was around 600 in day & 200-300 up close to a bright torch.

Make up gain is pretty good, might increase R6 a little & drop R19 to 270.
Would you measure the LDR while you play to get a value for R5?