News:

Forum may be experiencing issues.

Main Menu

Small clone for 1590a in SMD

Started by Rootz, September 13, 2016, 08:44:30 PM

Previous topic - Next topic

Rootz

Yet another development topic by me. Well, development... It's just miniaturising an existing pedal, this time the Small Clone. The only mod incorporated so far is a depth pot. I might add a vibrato or intensity switch.

Again I need some help. Could anyone of you point me to more info about converting this circuit to accept MN3207 BBD's? I'd like to make it easily switchable like the Pork Barrel. Can that be done here?



Rootz

Not much of a response here  :-[. I get that, it's not really groundbreaking either to just miniaturise an existing pedal.

However, I went ahead and converted the SC to MN3207. This meant redesigning the power supply for the digital parts, which isn't as simple as on a CE-2 where the Egg for the BBD is provided by the clock. I'll let the schematics speak for themselves when it comes to the other changes.

The first schematic is the most basic one with a simple zener regulator and intended for 9V use, with 9V max provided to the BBD. Should be safe for the BBD and in the optimum working range of the MN3207 (IF I understand the datasheet correctly; minimum THD at around 8V operation).



The second schematic is with a 7809 or 7805. Both should work, but when supplied 9V the 7809 will drop 2V and the BBD will run at a lower voltage.



The third schematic incorporates other changes as well. The digital components aswel as the bias for them runs of a low noise LDO regulator. The LT1761 has a max voltage drop of just 150mV when loaded as lightly as in this application. I ditched D2, which only provides a unnecessary voltage drop after Q1. Not needed as I used a NPN here and not an PNP. More true to the original would to use a PNP for Q1 and use D2. I just can't seem to find the reason why it was done that way.

The biggest change however is in the parts surrounding IC1. The MN3207 has a fair bit less dynamic range when compared to the MN3007. The SC by default boosts the high frequencies in IC1a by as much as 18 dB and compensates for this in IC1b. The Boss CE-3 is much tamer at this point and boosts the highs around 10 to 12 dB. To me that seems more reasonable and get some of the headroom back. All was simulated in LTSpice and by the end of the signal chain, the frequency characteristics between my tamed and a stock SC are the same. Apart from an obvious noise penalty, is there any reason to tame the signal to the BBD?


wgc

always the beautiful answer who asks a more beautiful question.
e.e. cummings

Rootz

#3
Thanks wgc. What do you think about lowering the signal to the BBD as I did? Any other improvements to be made, especially to the mods I did?

I made a quick mockup. The board looks like this:


dan.schumaker

That looks awesome, excited to see how all your 1590A SMD projects end up (and to hopefully build some of them  ;D)

Govmnt_Lacky

#5
Small suggestion....

Put the trimmer on the opposite side of the PCB. So it can be adjusted without having to remove the entire guts from the enclosure.

EDIT: OR..... make previsions to use a trimmer that can be adjusted from both sides of the PCB and have an access hole for a trimming screwdriver.  ;)

Rootz

Quote from: dan.schumaker on September 19, 2016, 02:43:51 PM
That looks awesome, excited to see how all your 1590A SMD projects end up (and to hopefully build some of them  ;D)

Thanks man! The boards for the micro sized Triple Wreck are on my way at the moment. Got an upgrade to swift service from the guys at OSHpark. Really nice. I want to see how that build is gonna work out. So that's the first one...

wgc

I'm not a huge theory guy, wish I knew more about bbd. But I can't fault your reasoning and the third approach is pretty interesting. Not sure what kind of noise you might get but to me a little grit might be cool in a flanger. Lowering the signal is an interesting approach too though personally I'm less inclined to do that vs make it loud and filter. :) 

Also not sure why a pnp would be used vs npn. Might be they got the parts cheap and it was a good way to save some $$. Or personal preference..?
always the beautiful answer who asks a more beautiful question.
e.e. cummings

Stomptown

Quote from: Govmnt_Lacky on September 19, 2016, 03:58:36 PM
Small suggestion....

Put the trimmer on the opposite side of the PCB. So it can be adjusted without having to remove the entire guts from the enclosure.

EDIT: OR..... make previsions to use a trimmer that can be adjusted from both sides of the PCB and have an access hole for a trimming screwdriver.  ;)

Good ideas here!

Rootz

That idea on the trimmer was clever! I should have thought of that and moved it to the other board side. Done that now.

I was thinking about lowering the signal because of the CE-3 from Boss. Much less signal to the BBD. Seemed to be a the way to go for a MN3207 with its lower dynamic range. Then I looked at the CS5 and CS9 from Ibanez. Both build around the MN3207 and in a very similar fashion to the others. The signal going to the BBD of the CS5/9 is at least as strong as in a stock Small Clone. Now those Ibanez chorusses are highly praised, right? I haven't read any mentions about noise and distortion.

The CS5/9 run the BBD regulated to 5 V. I think it's more logical to regulate the entire pedal to 9V instead, or regulate everything that is dc coupled to the BBD. The Pork Barrel has the BBD and clock regulated to 9.1V (zener) in case of using a MN3207, but the bias for the BBD is taken from the the main power supply. This means that when the pedal is powered from a 12V (for what ever reason you'd do that) supply the BBD still has 9V and is misbiased because the bias changed with the PS voltage. Not to mention you'd probably fry a dropping resistor and zener diode...

Conclusion: it's doable to fit a MN3207 in a Small Clone without altering the signal path to the BBD much (if my simulations and thinking are correct).

Govmnt_Lacky

Quote from: Rootz on September 20, 2016, 02:29:49 PM
That idea on the trimmer was clever! I should have thought of that and moved it to the other board side. Done that now.

You might want to also consider putting the BBD on the opposite side (socketed of course) It would make for easy repairs as that is the most likely fail item.

Quote
I was thinking about lowering the signal because of the CE-3 from Boss. Much less signal to the BBD. Seemed to be a the way to go for a MN3207 with its lower dynamic range. Then I looked at the CS5 and CS9 from Ibanez. Both build around the MN3207 and in a very similar fashion to the others. The signal going to the BBD of the CS5/9 is at least as strong as in a stock Small Clone. Now those Ibanez chorusses are highly praised, right? I haven't read any mentions about noise and distortion.

The CS5/9 run the BBD regulated to 5 V. I think it's more logical to regulate the entire pedal to 9V instead, or regulate everything that is dc coupled to the BBD. The Pork Barrel has the BBD and clock regulated to 9.1V (zener) in case of using a MN3207, but the bias for the BBD is taken from the the main power supply. This means that when the pedal is powered from a 12V (for what ever reason you'd do that) supply the BBD still has 9V and is misbiased because the bias changed with the PS voltage. Not to mention you'd probably fry a dropping resistor and zener diode...

Conclusion: it's doable to fit a MN3207 in a Small Clone without altering the signal path to the BBD much (if my simulations and thinking are correct).

Don't forget that all of the other pedals you mention used the companion MN3102 for clocking which could figure into the reasoning as to why they went with the 5V powering scheme. The SC used the 4047 capable of a higher voltage tolerance.

Food for thought... Good luck with this, it looks fantastic!  ;)

Rootz

Quote from: Govmnt_Lacky on September 20, 2016, 04:56:25 PM
Quote from: Rootz on September 20, 2016, 02:29:49 PM
That idea on the trimmer was clever! I should have thought of that and moved it to the other board side. Done that now.

You might want to also consider putting the BBD on the opposite side (socketed of course) It would make for easy repairs as that is the most likely fail item.

Quote
I was thinking about lowering the signal because of the CE-3 from Boss. Much less signal to the BBD. Seemed to be a the way to go for a MN3207 with its lower dynamic range. Then I looked at the CS5 and CS9 from Ibanez. Both build around the MN3207 and in a very similar fashion to the others. The signal going to the BBD of the CS5/9 is at least as strong as in a stock Small Clone. Now those Ibanez chorusses are highly praised, right? I haven't read any mentions about noise and distortion.

The CS5/9 run the BBD regulated to 5 V. I think it's more logical to regulate the entire pedal to 9V instead, or regulate everything that is dc coupled to the BBD. The Pork Barrel has the BBD and clock regulated to 9.1V (zener) in case of using a MN3207, but the bias for the BBD is taken from the the main power supply. This means that when the pedal is powered from a 12V (for what ever reason you'd do that) supply the BBD still has 9V and is misbiased because the bias changed with the PS voltage. Not to mention you'd probably fry a dropping resistor and zener diode...

Conclusion: it's doable to fit a MN3207 in a Small Clone without altering the signal path to the BBD much (if my simulations and thinking are correct).

Don't forget that all of the other pedals you mention used the companion MN3102 for clocking which could figure into the reasoning as to why they went with the 5V powering scheme. The SC used the 4047 capable of a higher voltage tolerance.

Food for thought... Good luck with this, it looks fantastic!  ;)
Socketing the BBD, that's something I'd really like! Moving it to the other side would mean I've to relocate it between the pots as there's an output jack on the opposite side. That would mean a pretty big overhaul...

Maybe I should start looking into daughter board and or other jacks (smaller ones like the ones from Zwee) for complex designs. How nice would it be to shoehorn a flanger/chorus/vibrato hybrid into a 1590a? Or just clip all legs when the BBD is dead.Doesn't matter, it's dead already haha.


Sent from my iPhone using Tapatalk

Scruffie

BBDs are not the most likely to die item as long as they're properly handled, while it should be socketed too, that crown actually goes to the 4047 and even they don't die that often.

The 4047 was chosen over the companion chip for its cheaper price and efficient driving only, it has nothing to do with voltages.

The 3207 is a lot less forgiving of bias shift than the 3007 hence the regulation, if you're only going to use the pedal with one power supply it doesn't matter but if you switch to one with slightly higher or lower voltage output the bias will be off.
The clock must see the same voltage as the BBD whatever happens and lowering the voltage of the entire pedal will effect the sweep range as the LFO can no longer go as high and low.

You could lift a 5V regulator with diodes up to 7.4V for a midway, remembering that the clock/BBD voltage buffer is dropping a diodes voltage.

BBDs 300X & 320X struggle mainly with bass frequencies which also tend to be the stronger p-p signal, the pre-emphasis reflects that.
Works at Lectric-FX

Rootz

A big thanks for that valuable information! Seems I was thinking in the right direction: take the bias for the BBD from the power that supplies it. Power CD4047 and BBD from the same point. Hence my (most recent) thoughts: either maximise the whole power supply to 9 volts via a LDO regulator or supply Q1, Q2, the BBD and the CD4047 from a regulated supply via a LDO. The latter gives the option to run the pedal as high a voltage as the other parts allow. The BBD will be safe and the bias will be steady. Last but not least, the LDO will have almost no drop when using a 9V supply.

Another option would be to regulate the clock, BBD, Q1 and Q2 (and the bias network for them) to 5V. Seems to be the sweet spot for the BBD according to some (or many, I don't know). Distortion in Q1 and Q2 will rapidly rise in this config, mainly 2nd and some 3rd. Not necessarily a bad thing... Is there a very good reason to run the aforementioned parts on 5V?

I always liked the idea of 9V over 5V better for the BBD and clock. For no apparent reasons whatsoever...

Brings me to the last point you note here: bass content to the BBD. In a CE-2 or CS9 e.g. the bass is greatly reduced by the 33n cap decoupling the first opamp to the pre-emphasis filter. Now that opamp is already unity gain at bass frequencies, the 33n cap drops even more lower bass (below 40 Hz IIRC). In that same place in the Small Clone is a 1u cap, which let everything pass (except DC off course). Other than that, I can't come up why a MN3207 would get more of a beating in a Small Clone than in a CS9 (5kHz at +18dB right before the BBD in the latter).

I'll post some simulations I made in LTspice, so you can better judge/see the ideas I had and the results of different filtering methods according to simulations. I know, simulations......

Scruffie

You'd have better luck giving the BBD a dedicated bias trimmer, it was just done through the 1st filter to save parts in the original design but you're going SMD so I doubt it'll cost you greatly. Bias Q1 with a 100k resistor to v.ref after the 1uF, add a 100nF after Q1 before the BBD and use your new 100k bias trim at the BBD to provide its Vgg voltage with a 6k8 resistor for close enough to 14/15th then add a cap at the output of the BBD with another biasing resistor, makes life a lot easier and gets you more headroom.

5V was chosen as the BBD operated happily off it and it meant a battery could be used in those old BOSS/Ibanez pedals, 8v2 is about the sweet spot for S/N & THD from past experience and from memory of the datasheets but they're good at 9V too and 5V is still fine just a little 'worse'.

A 3207 should be fine in there with the current set up.
Works at Lectric-FX