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

That would obviously be the better way to do the biasing; just like in the CS9, but with Small Clone values. CE-2 is done trough the first filter too. Any reason why the bias won't be stable when looking at the second image below? Anyway, I'm going back to the drawing board and see what I can come up with.

The LT1761 is adjustable and set for just above 8V. I saw in the datasheet that the sweet spot should be there ;-) and thank you for the confirmation!

All schematics in LTspice

LT1761 regulated Small clone

pre-emphasis outputs

pedal outputs

Rootz

Please take a look at this. Q1 and Q2 run of the normal VCC line again, the BBD and clock are powered from a separate regulated supply. The bias for the BBD is taken from that power supply aswel and placed right in front of the BBD. More parts, but a better design. Bias is further stabilised by a 10uF cap. This should be pretty solid. After the BBD there is a coupling cap and Q2 gets its bias from the regular power supply again. In theory it should now be possible to safely run the pedal at 9, 12, 15 or 18 volts without having to adjust anything. I upped the first filter cap to 220uF/25V. The BBD is always safe at 8V2, as is the clock.

You can socket the BBD. I just confirmed that the total height of IC + socket is 8.5 mm, so there is 1.5 mm room left between the BBD and the enclosure.

Who wants to actually see me build this thing?  ;D



Scruffie

Your new 10uF bias cap is forming an additional low pass filter and it's overkill anyway.

Otherwise, looks fine to me.
Works at Lectric-FX

wgc

Double dog dare you. 

Looks good to me.
always the beautiful answer who asks a more beautiful question.
e.e. cummings

Rootz

Ahhh the pressure wgc! This will be the final design prior to board fabrication I think. That cap is overkill, but it won't hurt either right? An extra 30 DB filtering in the bias, is that bad?
To you both a big thanks for having a look at the design. I'm confident that there aren't any fatal mistakes. Stay tuned to see how the Triple Wreck turns out. If the concept is good, this board will also go to OSHpark.


Sent from my iPhone using Tapatalk

brejna

Awesome layouts! I would love to buy some.. Especially if you attend to make Ce-2 :)
Also, are you using 0805 packages for resistors and caps?

Rootz

Thanks Brejna! Everything that works satisfactory will be put on sale. When the time is ripe. I wasn't really planning to do a CE-2/3 as there are already a number of others (Pork Barrel, Ensemble King and derivates, etc.) and I have got a Pork Barrel already. Yes, it's a bit big to my liking  ;D

The resistors are all 0603 as are most capacitors. Some caps are 0805, mainly the larger values. This way I can get a better voltage rating and usually a better capacitance/voltage behaviour (bigger ceramics seem to be better at that). Tantalums are either 3216 or 6032.

Rootz




Sent from my iPhone using Tapatalk

Rootz

This is off topic and consider this a work in slow progress. Both schematics show a CE-2 with minor mods to the power supply. The BBD and clock on their own voltage so you can run for example the BBD on 5V and the rest of the pedal on 18 Volts.
The first schematic shows a basic compander circuit. I still very much learning how to design and use a compander in a chorus. If i understand the working correctly, using one would greatly improve the noise figures and slightly the dynamic range. Does the CE-2 need it? Might not be. The Peavey CMC-1 compander chorus uses one and is praised for its pristine chorus. So why not try and learn something new. I saw in a couple of schematics that the first and last op amp in the circuit are just used as buffers. Is there no need for a high end emphasis anymore too when using a compander? I also noted that the r3 input is not always used on the compressor side. Why would one do that? Again, I am still learning very much about compander. A link to informative stuff (application notes already read) would be very helpful.


Sent from my iPhone using Tapatalk

Rootz

Oh on a side note: compander need at least 6v (sa571d to be specific), so the commanding chorus would obviously not work when the regulator is turned down to 5V. The other schematic will work.


Sent from my iPhone using Tapatalk

Scruffie

Total overkill in a CE-2 (and in most choruses) at the cost of extra parts, board space & dynamics, if anything was needed which it isn't as the CE-2 isn't a noisy chorus by any means then a simple gate like the CE-1 would suffice.

If you want 'pristine' chorus you may as well go digital anyway.
Works at Lectric-FX

Rootz

Valid points, thanks. Using that commander was really just out of curiosity.

Rootz

Back to the Small Clone clone after a very busy week. The idea of making it accept MN3007 and MN3207 BBD's crossed my mind again, so I just had to look into that idea better. As Vgg for the BBD is not taken from the clock chip, it wasn't as simple as in for example a CE-2 clone like the Pork Barrel. What I did is "flip" the Vgg supply around when you move the jumpers to select the chip. When you jumper for a MN3207 Vgg is 14/15 Vdd, when you jumper for MN3007 Vgg is 1/15 Vdd. At least in my brain this sounds logical. In a regular SC Vgg is grounded, but I can't figure out a way to do that with just two jumpers. Now the question: does the foregoing sound logical and does the updated schematic look okay to you guys? On a side note: all P points in the schematic connect to each other, as do the G points to each other. both jumpers are either horizontal (MN3007) or vertical (MN3207). The rest of the schematic is unaltered.


Scruffie

#28
Look at the Lectric FX - Celeste for how to flip VGG and power with just 2 jumpers, incorporating the bias trim/vgg set up I mentioned earlier.

You could also just flip it from ground (3007) to Vdd (3207) it's not imperative that it be used.

You don't want everything to share power and ground directly, you need to learn proper layout techniques, the clock is effectively digital.

Edit: Scratch that, the Celeste build document doesn't actually show it in action, anyway, you have your 68k & 4k7 resistors, if they're connected directly to the BBD power and ground pins, when you switch the jumpers they'll get switched too and if you invert the divider it does the job for you, just make sure the stabilizing cap doesn't switch and is always connected to ground.

Double scratch, on a reread of your post I see what you're asking, if you want to flip it from ground (as in the original SC) to 9v from switching for 3007 to 3207 just connect it directly to the corresponding power pin that is being flipped with the jumpers, so if Vgg is connected to pin 5 for the 3007 so ground, when the power and ground are flipped with the jumpers, it'll now be at 9V for the 3207 as it should be.
Works at Lectric-FX

Rootz

Thanks Scruffle, for your time and ideas again. You are a bit more knowledgable about this than me, so it seems.

Vgg is grounded in a regular Small Clone indeed, while the datasheet states it should be at 14/15 Vdd. I guessed both work like you say. I went for a simple solution, because it has a small foot print. Even 4mm trimmers are big on such a small board :o The Vgg stabilising cap is always grounded and doesn't switch like you say. What you mention after the double scratch is what I initially wanted. I'm afraid my post wasn't very clear... As both solutions work, either is fine to me. I probably shouldn't expect a difference in sound, right? I've read somewhere it might influence the gain of the BBD though... Decisions, decisions...

And you partly nailed it when you said I need to work on layout techniques. BBD and clock are on the same power (Vdd) which is separated from the main power through an LDO, but the BBD is not digital. I assumed this gives enough separation between Vcc and Vdd. Wrong? Ground is shared through the whole board. That's not the best way, but  I thought I'd get away with that and just use some massive ground planes instead. Clock is very close to it's own supply. Would you suggest to star ground the DGND at the AGND at the first filter cap or power inlet of the board? How does EH do this? Do they separate A and D grounds? So many questions, sorry Scruffle... I'll research a bit more about separation D and A supplies (Google is a close friend of mine).

To clarify where I'm in the design proces of the boards (I could probably lose 'a couple' of ground plane via's):

Top side

Bottom side