Powering and deticking the Polyphase

[Morocotopo]

Well, I took advantage of a few non working days to continue work on my Polyphase clone. It all started here:

http://www.diystompboxes.com/smfforum/index.php?topic=100570.0

I´m making a new thread so as to make following the development easier. Here´s my current schem:



First, the powering.

The idea is to NOT power it with a mains transformer as per the original but to be able to use a pedalboard power supply along with other pedals. Also, I have an Artec PS that has a max of 100 mA per output (all 9V), so ideally not use more current and voltage than that.
Two options: power it with 18V, a simple 15V regulator will do the job. Or, power it with 9V and use a charge pump. I went that route. As stated in the schem, an LT1054 works well. Close to it´s max current limit, but works. The circuit draws around 45-50 mA, so in voltage doubler mode the charge pump theoretically draws 100mA. Once I tried it, the actual draw is around 90 mA. Close! It didn´t burn or anything. I worry a bit about long term reliability, but... Should work with my PS. But there´s also the status LED´s... those will be powered directly from the 9V, I´ll use three (two for env/sweep switching and one for on/off). Will I take my PS over the edge? We´ll see.
The next snag was, I first used 1N4007 diodes for the charge pump but it didn´t reach 15V, too much forward V loss, and the LFO´s didn´t start. So I used 1N5819 schottkys, and it worked. But, it gave me around 16.2V at the output. That is, 18V minus 1.1V of loss in the pump plus the 0.35 or so loss from forward V in each diode. It works OK, I don´t know how well regulated the original power supply was... so next I tried to regulate the V with resistor/zener. No go, at the input, at the output, didn´t work, since I´m so close to my target current draw, the zener regulation needs more current to work... so I settled for using 16.2V for power. My other worry was, what happens if the PS gives more than 9V? I tried it with 10V and the charge pump gave 18V. Still worked OK, nothing burned, the only difference was a bit more "oomph" in the sound and the LFO´s got a little bit faster. So, I think I´m going with this power solution. Will do some more testing.
There´s another option, but I´m not sure how to implement it. The LT1054 datasheet mentions that the chip can do regulation. I assume they mean voltage regulation... that would be great, to get exactly 15V regardless of the input voltage, within reason. But the example they give is for a voltage inverter, that I don´t really understand. Could it be possible to have regulated voltage doubling? If anyone knows how to do it and can tell me, I could try that.

Well, that´s it for powering the thing. Next, deticking.

[Govmnt_Lacky]

Ariel,

I am really surprised that the LT1054 worked for you!

As far as how I interpretted the datasheet, the LT1054 has a MAX current supply of 50mA even in doubler mode. I somewhat proved this with a build I was doing that was drawing ~60-70mA when on regulated 12V supply. When I used the LT1054 in doubler mode via a 12V regulator, the circuit worked however, there was constant noise in the audio path because the pump was being overdriven.

[Morocotopo]

Gov, the datasheet says 100 mA OUTput current, but that´s in inverter mode I think. If that´s so, in doubler mode it´s 50 mA. Ohm´s law... double voltage, half the current. BUT if the thing can put out 100 mA in inverter mode, it has to source the same amount of current,100 mA... so max INput current 100 mA?
If your build draws 60 mA, the 1054 as a doubler needs 120 mA of input to make it. Over the top. And you have to add the regulator current.

In my case, yes, it works!

EDIT: By the way, the datasheet also says it can be paralleled. Have you tried that to get more current?

[Morocotopo]

OK, the deticking part.
As I already mentioned, the thing in the breadboard ticks, both LFO´s do. Of course I have leads flying everywhere, grounding not optimal... So here´s my proposed solutions:

-Decouple power in four branches with their associated parts with R/C filters (in red in the schem - A, B, C):
1)U6, Q1, Q2, Q3,  LFO
2)Audio path: U1, 2, 3, 4, 5
3)U9 LFO
4)U8, U7, Q5, Q4 Env gen

That should isolate the different parts from each other. also I think I´ll add 100n ceramic caps to all the audio opamps right at the power pins (not on the schem). Funny thing: the V supply oscillates at the lfo´s rate. That´s possibly the leds in the optos lighting up and down, drawing quite a bit of current in the process. It does it with every powering scheme I tried...But that´s not the ticking source, since it´s too slow.

Star the grounds in groups, marked 1, 2, 3, 4. Same criteria.

Some of it I tried on the breadboard, it helped some but to fully test it I´d have to rebuild the thing. I think I´ll design and build the PCB and try the mods directly there. Risky!!! But It should work...
The 4K7 and 0.47u also helped, but maybe they are redundant.

Well, that´s the plan. Any comment, please do!