| MT8816 VEE voltage [message #2599] |
Sun, 28 November 2010 04:14  |
 smrl
Messages: 23
Registered: February 2010
|
Junior Member |
|
|
So I've got a question regarding negative voltages to the MT8816 chip. I've gone through several chips right now, they seem to disintegrate after not too long. I'm sure I'm not giving them "safe" signals so this is quite likely part of the problem. I intend to put protection diodes on my in/outputs (Zeners? Schottkys?) but I've noticed something else. I've never quite gotten what I'd expect out of my negative voltage generator. I want to make sure that I'm getting maximum supply voltage so that I have the largest possible voltage compliance...
I'd expect if the voltage jumper is in the 5V position the output at Vout of the 7660 should be -5v, and the chip should see that plus the diode drop. If the jumper is at 9v, I should get -9, plus diode drop.
Now. Looking at the datasheet, it shows 13.2v as max supply voltage. (it also shows 16v as max. voltage for VCC?) The diode might make sense if the circuit were designed for maximum operation at +9, with 5v feeding the 7660 and consequently ~ -4.3v as VDD, giving 13.3/13.4v or so. This I could see working, however +/- 9V seems awfully out of spec.
That said, my negative rail has never been so low! Is there something in this design that I'm not understanding or should I have +/-9v? Perhaps I should check C100?
If I can't make sense of this, I've been considering building a lower-noise bipolar supply without a charge pump. Any suggestions for an ideal configuration there?
|
|
|
| Re: MT8816 VEE voltage [message #2601] |
Fri, 10 December 2010 14:41  |
dann
Messages: 829
Registered: December 2002
Location: Austin TX |
Senior Member |
|
|
Wow, we'll after looking closely at the datasheet, you're right that +/-9V is out of the operating range-- and this probably explains why you have some fried chips.... I missed that when I added the 9V jumper.
I haven't run a bend matrix at +/-9V for more than a few sessions, without seeing that my power supply is too hot and then switching back to +/-5V (all of the pre-built bend matrices ship at +/-5V, so no worries if you're reading this and bought one already built).  oops. I included that jumper as a last minute "Hey this should work, why not?". Well, thanks to your careful attention, now we know why not...
Looks like +/-8V would be the "Absolute Maximum" acceptable to the chip (Vdd=16V above Vee), although it would be out of the "recommended operating range" (Vdd=13.2V above Vee). That's a good idea to feed 5V to the 7660D and use the -5V to power Vee, then run +9V to Vdd. That puts us closer to the recommended operating range. A cut trace and a jumper wire would test this out...
As for power supplies... Yes, I wholeheartedly encourage the use of a true bipolar supply instead of a charge pump chip. Since designing the bend matrix I've used +/-12V switching supplies when I've needed a bi-polar supply and they are much more reliable than using a 7660D chip. A 8 or 9VAC adapter fed into a full-wave rectifier and then a 7808 and 7905 would do well... Or check these guys out, $1.75 each!!
http://www.hosfelt.com/contents/en-us/d220.html
(scroll down to part number 40-136)
By the way we are developing a matrix VCA using SSM VCA chips, that can handle +/-15V, as well as providing variable gain/attenuation (not just on/off). It's a lot more hardware but super more useful.
|
|
|
|
| Re: MT8816 VEE voltage [message #2609] |
Mon, 20 December 2010 09:13 |
dann
Messages: 829
Registered: December 2002
Location: Austin TX |
Senior Member |
|
|
Using these buttons (probably chinese equivlents for the production run):
mouser# 107-PS322RUG-EV
And yes, those are the power supplies i found from hosfelt electronics. good random surplus supplies
So the 4049 won't oscillate when it's connected to the 7660? Are you scoping it to determine that, or using a freq meter? What frequency are you getting? Sometimes I have to tune it down to 30kHz by adjusting the cap. I think the 7660 has a max freq of 40kHz (can't remember, too pre-coffee early morning to look it up). It'll generate a negative voltage without any frequency applied, but the internal frequency it generates is aroudn 10kHz, so you can hear it on the jacks.
In any case, I think a separate, non charge pump power supply is the best solution
|
|
|
Members
Search
FUDforum help
Register
Login
Home
E-mail to friend 
