Flintlock Flanger not Sweeping

[eh la bas ma]

I would struggle and find a way to heat several pads at once on IC4, to push the socket pins further inside the pads. You have to heat half of the pads all at once to make it move.

Moving quickly with the iron from pad to pad, reflowing the solder during one or two seconds on a pad and going to the next one, will work once they are hot enough .

The problem is that you'll have to push the socket with your other hand while you are heating all the pads, and the socket will get hot. You also need to be quick in order to avoid overheating the other IC near IC4.

I guess this could be a small part of your punishment, a taste of it, un avant-goût :

I will have to accept my punishment

I hope your punishment(s) will be exemplary, for your own sake of course, and in order to prevent other beginners from commiting the same mistake.
I can attest as a beginner, that I would think twice before doing the same mistake, if and only if your chastisement is really terrible and spectacular.

About the clicking noises, I would instinctively reflow and check all caps values, voltage ratings and orientations, especially the polarized caps. They all need to be rated above 27V (18x1.5).
If it doesn't work, I'd go on the hunt with the audio probe to find where it comes from, starting by probing the Mix trimmer.

[matopotato]

I would struggle and find a way to heat several pads at once on IC4, to push the socket pins further inside the pads. You have to heat half of the pads all at once to make it move.

Moving quickly with the iron from pad to pad, reflowing the solder during one or two seconds on a pad and going to the next one, will work once they are hot enough .

The problem is that you'll have to push the socket with your other hand while you are heating all the pads, and the socket will get hot. You also need to be quick in order to avoid overheating the other IC near IC4.

I guess this could be a small part of your punishment, a taste of it, un avant-goût :

I will have to accept my punishment

I hope your punishment(s) will be exemplary, for your own sake of course, and in order to prevent other beginners from commiting the same mistake.
I can attest as a beginner, that I would think twice before doing the same mistake, if and only if your chastisement is really terrible and spectacular.

About the clicking noises, I would instinctively reflow and check all caps values, voltage ratings and orientations, especially the polarized caps. They all need to be rated above 27V (18x1.5).
If it doesn't work, I'll go on the hunt with the audio probe to find where it comes from, starting by probing the Mix trimmer.

Thanks,
I think the socket itself is a problem, not how I soldered it in, so swapping for a fresh was my idea rather than reflowing/pusning deeper. It is quite deep as it is already I think.

I can check the caps but the circuit did work as it should at an earlier point, and I checked the kit before building, so if a cap has now failed is more likely, so the probe might find that.
Will both measure pins and probe now that the wobbling voltages are a bit more under control.

[eh la bas ma]

Good luck unsoldering your socket !

I can say the pcb is good quality and surely very resilient. I built a second unit because I really thought I destroyed the first one, and finally it's still perfectly good.

If you ever need to replace IC11 socket, i found a nice trick to make it desolder itself : shorting the LT1054 output pin to ground will start a melt down process. It will burn quickly, and the plastic socket will probably come off the board as a result. You also get a nice smell in the enclosure, as a bonus. This method has been tested and verified :

https://www.madbeanpedals.com/forum/index.php?topic=33571.msg322273#msg322273

(you'll find more voltage readings in this thread, as comparison)

[duck_arse]

I guess this could be a small part of your punishment, a taste of it, un avant-goût :

I will have to accept my punishment

I hope your punishment(s) will be exemplary, for your own sake of course, and in order to prevent other beginners from commiting the same mistake.
I can attest as a beginner, that I would think twice before doing the same mistake, if and only if your chastisement is really terrible and spectacular.

nah, youse gets no punishments. it is standard procedure to probe from easiest point to easiest point, because they are always best quality and not causing problems. but then when you find something and can't find any good clear reason, dig deeper. check more granular. down to the nitty AND the gritty.

also, when you inject 18V at the 18V point, all that 9V stuff is out, forget it, don't measure anything there. just forward, is all. that includes C33.

and IC sockets. don't try and save the socket, don't try and get it all out in a piece, destroy it [and the evidence]. cut as much of the plastic as you can, so that you are left with 8 pins separated, then remove each pin by one. this will result in less swearing and board damage.


my first boss used to say if the socket costs more than the IC, don't use a socket. that was a long time ago now. [better sockets.]

[matopotato]

I would appreciate some sound samples

It looks much better, but there're still some suspicious pads :

Checked them out and it seems to be some dirt, but I am reluctant to scrape it off to avoid damaging any traces. (Isopropyl did not help there)
Also the two places with this gunk stuff should actually connect, so had they been bridges it would only be additional contact.
The third place was fine.
Thanks for checking.

[matopotato]

nah, youse gets no punishments. it is standard procedure to probe from easiest point to easiest point, because they are always best quality and not causing problems. but then when you find something and can't find any good clear reason, dig deeper. check more granular. down to the nitty AND the gritty.

Good thanks, makes sense. Would not have come up with the idea to test at different point on the "same place" to find where breaches might happen.

also, when you inject 18V at the 18V point, all that 9V stuff is out, forget it, don't measure anything there. just forward, is all. that includes C33.

Yes, I was more trying to find my way. I tested to remove the IC11 and insert 18V at the D9 anode/C33 junction. But it just made sparks. I had read somewhere before that it might not be a problem so I quickly tried the pedal on/off but no sound either way.
Put the IC11 back in and went back to 9V and it behaves as previously, so the sparks were just for extra excitement.
I do not fully understand why it does not make the pedal run though. Had the same issues when I was dabbling with my trimpot construction, but now I had the 18V from the powerbank.

and IC sockets. don't try and save the socket, don't try and get it all out in a piece, destroy it [and the evidence]. cut as much of the plastic as you can, so that you are left with 8 pins separated, then remove each pin by one. this will result in less swearing and board damage.

Thanks, very good. Took a little while to get everything out, but worked well. A lot came off in the destruction phase so the leftover legs were quite small to get a grip on.
New socket is much better, and I doubt there is any fluctuations that should not be there.
Will do new measurements as well as probing.

my first boss used to say if the socket costs more than the IC, don't use a socket. that was a long time ago now. [better sockets.]

The ICs seem to still be ahead, but the precision sockets are closing in.
Very good point though.
Also, I probably have more spare sockets than spare ICs of the right type which also speaks for destroying sockets in my case.

[matopotato]

I managed to come by a few 4047s so I could put back the one to the phaser I borrowed from and a new to the Flintlock.
After lots of tweaking I managed to set the frequencies right (69.6k and 2.6M)
There is still some unwanted thump at the end of each cycle. I can dial it out but then the pots are only useful for some 50%-ish turn, so I doubt it is a solution. And at that point it does not sound like the sound samples I've listened to.
So I tried probing, with signal generator and guitar and looper, but they all seem to produce the thump almost anywhere I probe. Interestingly (?) IC1 pin1 did not thump, but pin 7 did. And the others that gave out any signal, be it quiet or raspy, still with the thump though.

So I went for another measurement session, that became two since the pot setting has impact.
First is with the pots fully CW or close to

V in: 9,21
TP "15V": 14,65-14,70
TP "7,5V": 6,98

IC1 (4558L)
7,00
7,02
6,38
0
7,25
7,31
7,30
13,95

IC2 (4558L)
4,70
4,71
4,68
0
6,97
6,99
6,99
13,97

IC3 (4558L)
7,07-7,21
6,96
6,94
0
6,99
7,18
6,83-7,35
14,34

IC4 (LM1458)
6,50-7,24       14,36-14,43
7,75             7,29-7,88
6,98-7,48        7,01
0                7,00

IC5 (CD4007)
2,63    14,70
2,56    14,70
0       0
2,60    8,00
2,62    0
0       0,27
0       2,45

IC6 (CD4047)
12,77-13,11    14,66-14,73
1,42-1,94      12,98-13,46
8,02            0
14,68-14,76    7,27-7,3?
14,64-14,76    7,20-7,29
14,64-14,76     0
0               0

IC7 (7815L) (not sure about the pinout so it varies from earlier posts, but values should indicate pins)
14,65-14,73
0
16,18-16,23

IC8 (4558L)
10,64
3,88
3,85
0
5,74
5,75
5,38
14,34

IC9 (CD4049)
14,72    0
7,20     7,20
7,20     7,20
7,20     0   
7,31     7,11
7,24     7,23
7,26     7,12
0        7,24

IC10 (MN3007)
14,55   4,95
7,21    4,99
7,31    7,13
0,99    0

IC11 (LT1054)
0       9,18
4,42    6,31
0       3,44
0       0

IC12 (4558L)
6,99
7,01
7,00
0
7,06
7,01
7,01
13,97

IC13 (4558L)
6,73
7,00
7,00
0
6,94
7,00
7,00
13,95

[matopotato]

The second set of measurements
This time with the pots all on full CCW

V in: 9,18
TP "15V": 15,07
TP "7,5V": 7,19

IC1 (4558L)
7,22
7,22
6,56
0
7,45
7,51
7,51
14,34

IC2 (4558L)
4,83
4,83
4,81
0
7,17
7,19
7,19
14,34

IC3 (4558L)
7,31
7,22
7,15
0
7,20
7,53
7,21
14,71

IC4 (LM1458)
2,14     14,87
8,00      3,7-12,3 **
*         7,20
0         7,19

*slow swing with jump: 3,?-7,80 then jump to 12,80. Then down to 7,? and another quantum leap to 3,10
**slow swing but no jumps

IC5 (CD4007) Many values drops ("->") down
2,1 -> 0,34     15,10
? ->   0,18     15,09
0               0,11
2 -> 0,57       8,25
0,47            6,83
0               0,02
0               1,99 -> 0,8 (slow)

IC6 (CD4047)
14,97     15,10
0,11      14,99
8,25      0
15,09     7,53
15,10     7,53
15,10     0
0         0

IC7 (7815L) (not sure about the pinout so it varies from earlier posts, but values should indicate pins)
15,10
0
16,58
(more stable this time)


IC8 (4558L)
2,24
3,97
3,96
0
5,89
5,90
6,83
14,71

IC9 (CD4049)
15,10    0
7,53     7,53
7,53     7,53
7,53     0   
7,53     7,52
7,53     7,53
7,54     7,52
0        7,53

IC10 (MN3007)
14,94   5,34
7,53    5,33
7,51    7,52
1,02    0

IC11 (LT1054)
0       9,18
4,43    6,30
0       3,45
0       0

IC12 (4558L)
7,20
7,20
7,20
0
7,20
7,20
7,19
14,35

IC13 (4558L)
6,94
7,20
7,19
0
7,14
7,21
7,20
14,35

[duck_arse]

I still don't like or understand [the entire bloody circuit] the voltages around IC8. I don't see how you can have 10V at output with 4V at inputs. the second set of measures is a little flaky looking too, I think.

explain again to me what happens at S C T - is S always connected to something. or just left open? IC2.1 produces "A" voltage, about 1/3 of whatever is VR. that voltage should not vary with pots, not MANUAL, not RANGE, not AUTO and not SPEED. whatever the 1/3VR is is fed to 5/6th's divider and then into IC8.1. that voltage at the (+) in should also not change.

the voltage at the T thing should not change, should be same as at 2.1 out. the voltage at C should be between 0V and 2.1 out volts, and will vary by ONLY the MANUAL pot position. IC8.1 takes the S voltage [whatever and however it gets that voltage] and gains it by -2, and adds that to the (+) in voltage, and outputs the result, feeding that end of the RANGE pot. being a low impedance souce, the opamp should be strong enough that that end of that pot DOES NOT vary with oscillations and what not at its other end.

check that section of voltages follows my guff. I'm only going by what they tell me.

IC4.2 is outputting oscillations. the AUTO pot wiper should vary with setting, between 68/(20+82) and (68+20)/82 ---- and yes, I don't care those numbers aren't right. whatever that voltage is will make the triangle wave output (at 4.2) slew sawtooth one way or other about syemmetric. now my spelling is gone as well. and the SPEED pot should make faster and slower. so the voltage at RANGE wiper should be the oscillations of 4.2 at one end and the offsetting volts at 8.1 at the other end of travel.

is any of this happening in your build?

[eh la bas ma]

So I tried probing, with signal generator and guitar and looper, but they all seem to produce the thump almost anywhere I probe. Interestingly (?) IC1 pin1 did not thump, but pin 7 did. And the others that gave out any signal, be it quiet or raspy, still with the thump though.

You have some unwanted noise right at the beginning of the circuit, starting exactly at IC1 ? R1 and R5 are clean ?


Pin1 is an output, so i would have thought that pin2 & 3 are clean too... You have noise at pin3 ?



Could be something wrong in this aera.

If not, I wonder if such noise, noticeable right from the start of the circuit, could be related to an issue in the power section ?

Edit :

If it's coming from the clock section :

"(...) setting the bottom of the sweep lower will result in a wider sweep but the sound may become too 'bouncy' "

In my experience, the sweep values from the instructions aren't always perfect for every Flintlock builds. For exemple, I lowered the highest point to 2.2 MHz instead of 2.6, in order to get rid of some background noises. I guess it is due to parts tolerance.

I wonder if increasing the lowest point of the sweep, let's say up to 80 KHz, or lowering the highest point has any impact on the noise ?

[matopotato]

I still don't like or understand [the entire bloody circuit] the voltages around IC8. I don't see how you can have 10V at output with 4V at inputs. the second set of measures is a little flaky looking too, I think.

Yes, I thought it odd that it would differ between the two measurement sets. I will redo the measurements. At least for IC8. Is there any other IC of interest based on the fault: "Thumping at the turn of a cycle, and not quite flange" ?

explain again to me what happens at S C T - is S always connected to something. or just left open?

The S C T is Sleeve, C = "Ring" and T = Tip. They are connected to an expression jack for a stereo 6,35mm plug that should go to an expression pedal.
So in the schematic I am assuming that how you press the pedal the more you allow the MANUAL to be included in the circuit, without a pedal it seems the MANUAL does not do much, which corresponds to turning it does not change anything that I can notice.
So I added an expression pedal but it did not change anything. Regardless of where I put the MANUAL pot.
So I thought the expression pedal might be of the wrong value or something, so I shorted the T and C, and T and S and S and C respectively to mimic a full or minimal shift on the expression pedal. And in one combo the thumping went almost quiet. But only with MANUAL fully CW. No noticeable change to the Flange sound though.

IC2.1 produces "A" voltage, about 1/3 of whatever is VR. that voltage should not vary with pots, not MANUAL, not RANGE, not AUTO and not SPEED. whatever the 1/3VR is is fed to 5/6th's divider and then into IC8.1. that voltage at the (+) in should also not change.

the voltage at the T thing should not change, should be same as at 2.1 out. the voltage at C should be between 0V and 2.1 out volts, and will vary by ONLY the MANUAL pot position.

And what if I connect T to C or C to S? C-S ought to change the total R for IC8 and thus change its result?
T to C I do not understand. Perhaps T C and S are all connected to some extent when the expression pedal is connected...
MANUAL is supposed to set the "delay time range", but like I wrote earlier, nothing I can detect. Perhaps if I can get ti to flange and lose the thump that might be clearer.

IC8.1 takes the S voltage [whatever and however it gets that voltage] and gains it by -2, and adds that to the (+) in voltage, and outputs the result, feeding that end of the RANGE pot. being a low impedance souce, the opamp should be strong enough that that end of that pot DOES NOT vary with oscillations and what not at its other end.

check that section of voltages follows my guff. I'm only going by what they tell me.

OK, will re-measure IC 2 and IC8

IC4.2 is outputting oscillations. the AUTO pot wiper should vary with setting, between 68/(20+82) and (68+20)/82 ---- and yes, I don't care those numbers aren't right. whatever that voltage is will make the triangle wave output (at 4.2) slew sawtooth one way or other about syemmetric. now my spelling is gone as well. and the SPEED pot should make faster and slower. so the voltage at RANGE wiper should be the oscillations of 4.2 at one end and the offsetting volts at 8.1 at the other end of travel.

is any of this happening in your build?

SPEED does change speed now. And RANGE does change something I could interpret as range, and effect that feels like a wider sweep at CW as opposed to narrow sweep at CCW.
I can re-measure with the RANGE in different positions and the SPEED at different settings (as long as it does not fluctuate, those readings are frustrating to make. So I guess that will be the case here)

I tried in some act of desperation coupled with ignorance to cycle around the 4558s among themselves, just in case I had broken anything. Just one shift through, but no effect. Not saying they are all good, but it did not affect the overall output so I doubt they are the main source for any error.

I will be able to buy 4007, 4049 and 78L15 tomorrow, also hoping for a 2N5457. It is not expensive so might as well give it a shot.
The 3007 and 1054 would have to wait more proof before trying to swap them.
I did fry a 4047, so something else might have gotten its fair share of that same voltage spike, but it is a bit of a long shot I guess.

Now that there is no wobbling in R on VB and the measurements are at least stable in nature compared to earlier when the 1458 was sitting loose in the socket, should I still try such things as:

• 1458 -> TL072?
• Remove 1054 and run 18V in the 9V input?
• Also remove the 78L15 and inject 15V? Although I do not know how I should go by producing 15V. I have 9 and 18 to work with, but no 15V adapter as none of my other  pedals run on it.

[matopotato]

So I tried probing, with signal generator and guitar and looper, but they all seem to produce the thump almost anywhere I probe. Interestingly (?) IC1 pin1 did not thump, but pin 7 did. And the others that gave out any signal, be it quiet or raspy, still with the thump though.

You have some unwanted noise right at the beginning of the circuit, starting exactly at IC1 ? R1 and R5 are clean ?


Pin1 is an output, so i would have thought that pin2 & 3 are clean too... You have noise at pin3 ?



Could be something wrong in this aera.

If not, I wonder if it could be related to the power section ?

Edit :

"(...) setting the bottom of the sweep lower will result in a wider sweep but the sound may become too 'bouncy' "

In my experience, the sweep values from the instructions aren't always perfect for every Flintlock builds. For exemple, I lowered the highest point to 2.2 MHz instead of 2.6, in order to get rid of some background noises. I guess it is due to parts tolerance.

I wonder if increasing the lowest point of the sweep, let's say up to 80 KHz, or lowering the highest point has any impact on the noise ?

Thanks,




Pins 1, 2 and 3 did not have the thump noise. Pin 7 did.

Also checked the diodes, and although they are in circuit, I don't get any readings that would indicate either of them failing, values seem to make sense in respective context.

[eh la bas ma]

If the noise appears at pin 7, you might want to check everything connected to pin 7 : C2, R2, everything up to R11 in this aera. Also the R14, R30, etc. aera. I would check the soldering and make sure there are no mistakes in values.

I really think you should put a jumper as instructed between C & S, and forget about the expression jack, until you have the circuit fully working.

Manual is supposed to be very noticeable when you turn it. You can do a manual flange effect just by turning it back and forth.
If it's set in the first half of the rotation, the flanger will work on low frequencies. In the second half, you get a cleaner, brighter modulation.

[matopotato]

If the noise appears at pin 7, you might want to check everything connected to pin 7 : C2, R2, everything up to R11 in this aera. Also the R14, R30, etc. aera. I would check the soldering and make sure there are no mistakes in values.

I really think you should put a jumper as instructed between C & S, and forget about the expression jack until you have the circuit fully working.

Manual is supposed to be very noticeable when you turn it. You can do a manual flange effect just by turning it back and forth.
If it's set in the first half of the rotation, the flanger will work on low frequencies. In the second half, you get a cleaner, brighter modulation.

All good points thanks,
Will bridge C&S, actually it is a switching jack, so they are already connected... makes sense then. But then the MANUAL pot is doing next to nothing as of now. Perhaps another area of problem.
Will also check your suggestions above.

[ElectricDruid]

Manual is supposed to be very noticeable when you turn it. You can do a manual flange effect just by turning it back and forth.
If it's set in the first half of the rotation, the flanger will work on low frequencies. In the second half, you get a cleaner, brighter modulation.

I don't know how it works in this circuit exactly, but in some, there's an interaction between Depth and Manual. If you've got the Depth turned up to full, there's nothing left for the Manual to do. Try turning Depth down and see if you start to hear something happening when you move Manual. If that still doesn't do it, there's definitely something wrong there somewhere.

The other thing that I noticed is that there's been a lot of talk about setting up the clock frequencies, but how is the BBD bias set up? You can make things thump by having that a bit out sometimes too. Just a thought.

[matopotato]

Manual is supposed to be very noticeable when you turn it. You can do a manual flange effect just by turning it back and forth.
If it's set in the first half of the rotation, the flanger will work on low frequencies. In the second half, you get a cleaner, brighter modulation.

I don't know how it works in this circuit exactly, but in some, there's an interaction between Depth and Manual. If you've got the Depth turned up to full, there's nothing left for the Manual to do. Try turning Depth down and see if you start to hear something happening when you move Manual. If that still doesn't do it, there's definitely something wrong there somewhere.

Thanks, I wasn't sure how it worked. Will try it out.

The other thing that I noticed is that there's been a lot of talk about setting up the clock frequencies, but how is the BBD bias set up? You can make things thump by having that a bit out sometimes too. Just a thought.

I have a few trimmers: Mix, Bias, Auto, T-Enhance and Vol
I will try with Bias and Auto a bit more, but I remember trying all of them earliier with some effect, but not sorting out the base sound.
The Bias is quite "early" in the path and close to the IC1 pin 5 & 8 which could affect its output.

I also now realize that the Mix is probably how much of the clean signal to mix with the modulated signal before reaching the output, which then means that before ("to the left") of the Mix trim, the signal ought to be clean.
Then there should be no flangish sound when probing around in that part. I will have to check the schematic details, but I am guessing it should be possible to figure out where the signal ought to remain clean.
And if the thumping is in those areas, I would suspect power. If not, then possibly either of the remaining ICs or surrounding components.

[duck_arse]

explain again to me what happens at S C T - is S always connected to something. or just left open?

And what if I connect T to C or C to S? C-S ought to change the total R for IC8 and thus change its result?
T to C I do not understand. Perhaps T C and S are all connected to some extent when the expression pedal is connected...
MANUAL is supposed to set the "delay time range", but like I wrote earlier, nothing I can detect. Perhaps if I can get ti to flange and lose the thump that might be clearer.

this. the very purpose of a circuit diagram is to explain the circuit connections. how hard is it for the draughter to include the jack switching arrangement? it does make sense for TSC to be some sort of shorting jack, but that's not for us to guess at. grrrrrr.

[matopotato]

With less than full Range, Manual has an effect. Thanks @ElectricDruid.
The Bias trimmer has effect on the thumping, but does not dial it away, but rather change character between clicky and thumpy.
I still think I should be able to find out what causes the thumping problem, and then hopefully the sound will also be more "flangy".

I manage to buy one CD4007, one CD4049 that I tried in, but no change.
I also tried a new 78L15, but no change either.
Then I was going to try changing the 2N5457, but it is a bit more costly in my shop for a random swap test, so i googled that (among others) BF244A might be a replacement, and that was cheap enough to try.
And now the thumping went almost all away, but quite a murkier sound. So clearly not a 1-1 replacement in this case.
I got an SMD 2N5457 and I have some adapter I will try later on, or just buy the TO-92 one. So a bit more hopeful, although it might well turn out to be the same problem with a proper swap of the Q1.
Strangely, the pinout says DSG, but my meter has it GDS and that would not have fitted on the PCB without twisting some legs around in either prepared slot.
And it seems to work as the standard pinout (DSG).

[matopotato]

explain again to me what happens at S C T - is S always connected to something. or just left open?

And what if I connect T to C or C to S? C-S ought to change the total R for IC8 and thus change its result?
T to C I do not understand. Perhaps T C and S are all connected to some extent when the expression pedal is connected...
MANUAL is supposed to set the "delay time range", but like I wrote earlier, nothing I can detect. Perhaps if I can get ti to flange and lose the thump that might be clearer.

this. the very purpose of a circuit diagram is to explain the circuit connections. how hard is it for the draughter to include the jack switching arrangement? it does make sense for TSC to be some sort of shorting jack, but that's not for us to guess at. grrrrrr.

The switching jack


The Tiger colored goes to "T". It will go to the Tip of the expression pedal if engaged, else it does not go anywhere
The redBlack goes to S (=Svart, Swedish for Black)
And the Candy colored to C (red-white). So C and S are connected by default.
With an expression pedal C and S are disconnected directly and the expression pedal will provide varying R I think.
Black wire is to ground and sleeve of expression jack.

I am truly sorry for not explaining this better earlier and for the frustration this caused.

[duck_arse]

thank you for that photo, I wish we had it ages ago. but I wasn't grrrring you, it was for the schematic drawing person.

for the jfets we use, drain and source pins are interchangeable. just don't mix with gate is all.

it seems an odd choice to have the oscillator bias voltage (VB) the same bias as nearly all the audio section. if you were very adventurous and good at breadboarding, you could maybe dis/prove the bias/thumping business. you could mabye butcher a spare opamp and lift legs 5,6,7 from the socket so IC2.2 is out of circuit. then breader a separate voltage divider and buffer and jamb that into the socket pin 7.

hmmm, nice try, but that still puts VB to the bloody oscillator. might be better to lift IC4.2 (+) input out of socket and feed that pin its own clean reistor divided and cap smoothed ref volts.


and - I wonder why no smoothing cap on the divider at R37/R38. you could try tacking on a 4u7 or 10uF or so.