Uh oh. I'm trying to hold with ya here Harry. ;) Honest. But I've been scratching my head a time or two ...... so let's see where I get to. :D ;)<br><br><div><span class="gmail_quote">On 12/26/05, <b class="gmail_sendername">
harrybissell</b> <<a href="mailto:harrybissell@prodigy.net">harrybissell@prodigy.net</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Howdy Ryk...</blockquote><div><br>Hey sir. :) Again - many thank yous for your help. :) <br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I would not use the LM329 wholesale. If you want to power a whole<br>module...<br>a shunt regulator makes less sense.</blockquote><div><br>But - if I had a simple voltage follower, after it - could I not rely on THAT, for the required current supply for everything in the module? There-by, not putting too much strain on the LM329?
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Shunt regulator basically works by eating whatever excess current is<br>available.
</blockquote><div><br>Eating the excess CURRENT? Uh oh. I was under the impression that they acted like a normal Zener diode, eating the excess voltage. (majorly, scratching my head) :O (Very - LM329 clued out - except for limiting the amount of current that can be pulled through it, with a resistor between it and the Vsource.)
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">So if you need a lot of current, you use a smaller series resistor, and<br>the 329 has to eat more, gets hotter, wastes more power etc....
</blockquote><div><br>I haven't totalled the # of CHMOS chips that I'll have in this module of the panel overall. (It'll be multiple modules, instead of 1 PCB overall.)<br>I'll roughly guess it between 4 and 8. (Should I first check individual CMOS chip current drains, and total them?)
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">What I usually do for CMOS inputs... is use a series resistor with a<br>diode to Vdd and a diode to Vss. Maybe 50K resistor is good... choice between drive current and not loading the source too much. If the voltage goes above Vdd or below Vss it is clamped.
</blockquote><div><br>Ya know? I finally stopped scratching my head here, thankfully. :) I remember reading about this for protection - probably in my CMOS Cookbook. I'll have to reread. My original headscratching on THIS one, was hooking the diode/resistor pairs, between the +V power input to the CMOS chip and the Ground pin of the chip to ....... never mind. ;) (
I.E. I was missing what you were meaning. :O Doh.)<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">You can even use like 100K resistor and LET the input diodes of the CMOS clamp it for you. Downside is... this might power up the CMOS, maybe you didn't want that (probably you won't care most times....)
</blockquote><div><br>Now here - where I'm minorly scratching my head - is this considered - if the CMOS chip is drawing more current than can logically be supplied through the 100k to it? Thereby heating either or both?
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Running the 3080 Gm bias input from a resistor is usually a loser. Its<br>good for a very limited range of control. What is better is running it with a current.
</blockquote><div><br>But - I was under the impression that the current into pin 5 of the 3080, HAD to be limited - hence the resistor. ??? (Major head scratching.)<br>(And in my prototyping as mentioned above - from this afternoon - I was having a max. 10:1 control range.)
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">It can be as simple as a single PNP tranny. Another way is to use one<br>opamp and one transistor... this gives quite good control. A third way
<br>is a pair of PNP trannies (can control over V+ to V- range). All of these would be temperature sensitive unless you use tempcos (if you need that much stability). For a sequencer, probably not!</blockquote><div><br>
One of my next steps, WAS going with a single tranny / opamp current 'source' for the 3080.<br>Now - if I recently recall correctly - I COULD go with a matched pair of trannies source, and possibly get within 2 to 3 volts of 1V/Octave accuracy, with a tempco? (Using 2N3906s and one of the SMD 1K tempcos that I bought a bunch of. I would LIKE to have frequency doubling possibilities for 2 to 4 octaves if I could.)
<br>Also - I was thinking of looking towards +/-V input possibilities from the external world, so as not having to worry about protecting things from -Vs coming into it ........<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
With a resistor for Gm bias, you need to vary voltage. Try getting a<br>1000:1 range that way and you need voltages WAY outside the supply. Use current, and you can do 10,000:1 if you want to. The idea is, the current source changes its impedance to suit the current you want. At uA levels its like having a high megohm resistor... at mA levels it like having a 10K resistor.
</blockquote><div><br>Now - I'm sorry, but theory wise - this IS way past me, with basically no experience in design. :( I'm grasping though, if I'm not mistook - by understanding that I have to have a CURRENT source (via the tranny side of things, supplying a constant current - no?) rather than a Voltage source, as I was during today's prototyping.
<br>No? (For as far as I've typed - that I currently (pun intended as well. ;) ) understand. :) )<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Good ways can be found at Jorgen Bergfors' site (google Bergfotron) and look under the VCA, then "VCA shootout" for schematics of OTA VCAs </blockquote><div><br>Done and bookmarked. :) I also have a fair amount of schematics with current sources in them, in Electronotes; Barry's book and bookmarks. :)
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">The high megohm post are a loser. Use reasonable values to get voltage (maybe 10K to 50K) and then drive current sources with that.
</blockquote><div><br>What I have gotten lost on here, is the very first sentence - are you meaning a megohm range resistor, after the 100k pot from the Voltage source? I.E. +15V to 100K (to Vlimit it) to 100k pot to 1meg to pin 5 of the 3080?
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">If you want to regulate power supplies... use 78xx / 79xx regulators (or<br>LM317 / 337 for adjustable). Then you could still use the LM329 for precision needs.
</blockquote><div><br>I DO have various values in the 78xx series, from when I was building my own power supplies, in the earliest days.<br>Now - If I'm not mistook - I could hang a pair of these off of my main power supply, for keeping the voltages constant? ( I could go this route, say with 8 or 9V Vregs in the 7800/7900 series. I DO believe I have 2 x LM317's still as well. Shall have to check).
<br>Ultimately - I DO want a stable voltage supply for my Master Clock, as it will be driving most of, if not all, of my sequencers that will be going into this panel. Seeing minor variation in frequences today in prototyping, was a little uncomfy. :(
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Another good way to protect against input onervoltage is a simple zener diode.
<br>A 5V zener will clamp at 5V and -.7V. Just remember to have a series<br>resistor in front of it. You might want to use like 6.2V on a 5V signal so you don't clamp until AFTER you pass the highest input voltage.<br><br>
H^) harry</blockquote><div><br>I do have a bunch of 5.1V zeners and 9.1V zeners that I bought from 'write off' years ago. :) And I DO wish to protect from overvoltage as well ...... I was going to say on the Clock CV input, but - I will be having Inputs for quite a few portions within this panel. I hadn't yet given thought to protecting them all.
<br>Now is the time, to definately decide a Voltage standard for interfacing. (Have made my 3 x ADSRs +5V out. I may stick at that with everything. +/-5V.)<br><br>Harry. :) Many, many thanks for your help! Yourself and Magnus, shall be properly credited, when the schematics for the final designs go up on my webpage, for this panel of modules. :) :)
<br></div></div><br>-- <br>Take care,<br>Warmth and Peace,<br>Ryk<br><br><a href="http://deathlehem.bravehost.com/damian.html">http://deathlehem.bravehost.com/damian.html</a> - D.A.M.I.A.N.;s webpage - Dec 2105 update