| Electronics > Projects, Designs, and Technical Stuff |
| Please critique my charge pump voltage inverter design |
| (1/1) |
| wnorcott:
Hello friends, I am using 4 of the inverters of a CD40106 Schmitt trigger hex inverter as the clock of a charge pump voltage inverter in this design,see snippet of the schematic. The remaining two gates are used for something else. Anyway using the charge pump to get -9V from a 9v rail to power a CD4053 which when used for switching analog signals, needs a bipolar supply. Is what is shown better than a ICL7660 and should I increase the switching frequency? Looking at it on the scope Rigol DS1054Z, the switching frequency of the oscillator 21.4 Khz was accurate and the square waves looked square and after later soldering in the charge pump section the negative voltage looked like DC and it was powering up the 4053 but was not putting any audio through it. When I tried the listening test off the scope the switched audio sounds fine but, just a lingering doubt. I heard bad things about the ICL7660 but I happen have a tube with about 20 left so I could change over. The 40106 plus the passive components is still quite a bit cheaper and space on the board is not an issue. Thanks. I have a bunch of the 40106 that were only $0.25 each while the ICL 7660 are closer to $2. Everything is through hole components. |
| Ian.M:
Paralleling Schmidt trigger input gates is always a bit dodgy, because they are unlikely to switch at exactly the same time. If the input transitions fast enough, you'll get away with it, but you should check the supply current at the moment of transition with no load. I'd definitely want the switching frequency to be a lot higher than double the highest audio frequency your circuit will pass so that any intermodulation products are well above the audio range. Around 50KHz should be suitable. |
| Zero999:
The TC7660 is cheaper. Is this really necessary? What's the maximum voltage swing of the audio signal? Can't you simply bias the CD4053 at half the supply voltage and AC couple the audio signal? There are also analogue switched available which can control negative signals, without an external negative supply, as they have the charge pump built-in, but they're more expensive. |
| wnorcott:
Thanks Ian I will bump it to 50 KHz switching rate Zero999 I checked, and you're right. I can get ICL7660 much cheaper than I thought they were, even on eBay I saw them for 10 for $USD 1.29 shipped. I could use the biasing at 1/2 with the coupling caps. Maybe that is better. The signals being switched would not be more than 9 volts peak to peak anyway. Maybe I am overthinking it having a bipolar supply for the 4053. I was trying to treat it as a true switch to select either wet or dry audio signal and the wet is a preamp that already is AC coupled with its own cap so it seemed like I was adding a bunch of extra caps everywhere just for the switch, if I did the bias it and AC couple it approach. Thanks, all. |
| Zero999:
Yes, I think it's much easier and cheaper to AC couple. A charge pump requires two extra capacitors anyway, then there's the extra board space and cost for the IC. An analogue switch can never be treated as a pair of dry contacts because it doesn't behave as such. The voltage range is limited by the supply and the on resistance changes, depending on the signal level, which will introduce distortion, if the load impedance is low. You might be able to use the existing capacitor in the pre-amp but if it's polarised, it might need to be reversed. |
| Navigation |
| Message Index |