Egodriver Double voltager

[teknoman2]

Also I took out the max and measure the switch pins,

in every position pin1 is the same 8.6v
but on pin3 is 8.1v and 7.4v

So I believe something wrong is on this side of circuit.

From what you said the normal are pin1 and pin3 voltage of switch must not change.

Am I right?

[RobA]

I just double check again my voltages and I found that something odd,

I measure ground voltage 0.03-0.1V

Is that normal?

Yeah pretty normally. It rarely will read at exactly zero and it'll fluctuate around a bit too.

[RobA]

Also I took out the max and measure the switch pins,

in every position pin1 is the same 8.6v
but on pin3 is 8.1v and 7.4v

So I believe something wrong is on this side of circuit.

From what you said the normal are pin1 and pin3 voltage of switch must not change.

Am I right?

Yes, I think you are right. The voltages on pins 1 and 3 shouldn't change with switch position. With the IC out, pin 1 should be a couple of diode drops below pin 3 and should stay the same no matter the switch position. I still think that this points to a switch problem or a solder bridge either at the switch or the PCB. Some of the other numbers with the IC look strange, but those may all disappear when you get the switch issue fixed.

One other thing that could be consistent with the problem here though is to make sure that you have D3 and D4 in right. I don't think that's the issue, but it's worth checking.

[teknoman2]

Personally I believe that isn't a switch problem,

I measured on D4 positive pin and on C13 positive pin is 7.6v.

So its normal on switch pin to have the same voltage.

I think that D3 and D4 (1n4001) are damaged.

Do you think worth changing them is a good idea?

[RobA]

It could be that there is a problem with them. If you have a diode setting on your multimeter, you should be able to check them in the circuit. I just tested it on my breadboard in the circuit and it's a tiny bit off, but does show the diode drop pretty close.

Just so you'll know, the MAX1044 did really well in the noise/oscillator frequency test. It was at 47kHz with the boost pin on. So, way outside the audio region.

I have to go get some sleep. Good luck tracking it down. I'll check back in when I wake up.

[teknoman2]

Thanx mate I will keep you posted if anything change,

Thanx again.

[teknoman2]

Finally I managed to find my issue but we dont know the cause.

For some mysterious reason when I use max 1044 I measure grater voltage drop across the resistor.

Thats why we have different voltages on the switch.

Now I just lift up 1044 pin1,

So when switch is on we have:

10mA voltage drop
8.8v battery and 7.8v across the 100R and 14v on pin 8 of TL082.
switch lungs 7.8v, 14v, 14v.

When the switch is off :
5mA voltage drop
8.8v battery and 8.5v across the 100R and 8.5v on pin 8of TL082.
switch lungs 8.5v, 8.5v, 16v

[RobA]

That all does make sense. (I was reading on and off for the switch backwards last night -- had me confused.) The efficiency of the charge pump is reduced when you are in boost mode. So, it drops more voltage when it draws more current in boost mode than it does in normal operation with pin 1 disconnected. The problem is that the MAX1044 is way down in the audible range if you don't use the boost pin.

Also, I don't think it should be dropping that much voltage for the amount of current that circuit should pull. Could be that there is a short somewhere that is drawing too much current when the charge pump is being used. I doubt that. It could be that the charge pump is damaged. It should be able to supply about 20mA and at least 10mA in boost mode without that much drop.

The types of caps you use can have a pretty big impact on the charge pump. Using a low ESR type is best and I did some experiments that showed up to a 10% difference with some caps. Moving the caps up in value can help, but too big can hurt. Paralleling a big cap with a small might help.

Going to a smaller resistor on the input will lower the voltage drop there a bit, 10Ω-20Ω might be good. Using 1N5817's for the diodes will get you back quite a bit of voltage too.

I'm going to breadboard the rest of the circuit to see what I get for the current drop in both modes.

[RobA]

OK. I bread bordered the whole schematic. It sounds good.

At first I had it with the 100Ω and 1N4001's. The voltage numbers were very similar to the numbers you last posted.

Then I moved to a 10Ω and that alone makes things much better. After putting in the 1N5817's, things improved more. With the 10Ω and the 1N5817 diodes these are the numbers for currents and voltages;
Switch On:
   Current: 8mA
   9.19V
   17.74V
   17.74V
Switch Off:
   Current: 4.3mA
   9.13V
   9.13V
   18.10V

I tried with the LT1054 as well and it gets a tiny bit more voltage, but I don't think it's really worth it.

I suspect something is going on between the input resistor and the input impedance of the charge pump and the 100Ω is high enough that it is working like a voltage divider.

I'd give it a try just changing the resistor and see what you get. It's definitely worth moving to the 1N5817's too, but the 10Ω resistor seems to be the first order fix. You might even be able to get by without a resistor there at all.

Your design works well for me with that one fix.

Edit: I would leave LV pin disconnected too so that it can work with other charge pumps as well as the MAX. I had it off in all my testing.

[teknoman2]

I just bought some 1n5817 and by the end of the day I will make these changes.

I will post my results.

[teknoman2]

I made the changes as you suggested,
100Ω to 10Ω
all diodes 1n5817
change max1044 to icl7660 (they have the same datasheet)
left open Lv pin6.

These are my measurements.
Switch On:
  8.58V
  17.1V
  17.1V
Switch Off:
  8.9V
  17.9V
  17.9V

every works fine now...
thanx again for your help and your time.

please PM me your address to send you one board for your help.

[RobA]

Excellent. Congrats on the design and your board working. I'll PM you my address. It'll be cool to build this one.

I did a bit of playing around with the charge pumps when I had it on the breadboard. I found the Microchip TC7660H gave me the best results, highest voltage combined with highest clock frequency and least current draw from the charge pump. They are pretty cheap too, about $1.20US from Mouser I think.

[teknoman2]

Thanx again for your help,
I am going to make an order to mouser
including both tc1044 and tc7660h.
Since we managed to solve it,
I jump to the 'mangler' using a charge pump for negative supply.

If you want I can PM you the schematic for this one as well.

[teknoman2]

I would like to ask you one last thing,

Why we have to connect Boost with V (pin1 with pin8).

From a brief looking none of the datasheets indicates that we have to do this.

Do you know the reason of this connection?

[teknoman2]

I believe I found it,

if we connect Boost with V+ we increase the oscillation frequency from 10kHz to 45-47kHz.
which is out of range of guitar frequency.