Technical question about good low power rail to rail opamp
in [Design]
|
Prev: Could lightning flashes damage a camcorder ?
Next: Looking for inductor 200uH/1.5A, under 0.400" footprint
From: speff on 12 Apr 2010 22:16 On Apr 12, 7:33 am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: > Technical question about good low power rail to rail opamp. ^^^^^^^^^^^^^^^^^^^^ Please, JP, this is primarily a political discussion group. > I published 'scope_pic' > http://panteltje.com/panteltje/pic/scope_pic/ > and am getting questions as 'to where to connect the input etc..'. > > The question is how to make a 1 MOhm, few pF, input with a bandwidth of about 1MHz, > with as few components as possible, that also creates a +2V DC offset > (to get the ADC midrange and allow negative signals), Are you looking for gain of 1, or do you need other gains too (like a real 'scope)? The GBW of the amplifier will have to be much higher if you want to see a 1mV 1MHz signal fill the display. > So that leaves opamps, and the ones I found (5V rail to rail) > have either low bandwidth, bad slew rates, > incredible distortion, or all of those at the same time. I guess you could use the +/-8V supplies and avoid the need for a R-R amp?
From: MooseFET on 12 Apr 2010 22:39 On Apr 12, 4:33 am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: > Technical question about good low power rail to rail opamp. > I published 'scope_pic' > http://panteltje.com/panteltje/pic/scope_pic/ > and am getting questions as 'to where to connect the input etc..'. > > The question is how to make a 1 MOhm, few pF, input with a bandwidth of about 1MHz, > with as few components as possible, that also creates a +2V DC offset > (to get the ADC midrange and allow negative signals), Op-amp idea: The LT1352 gives a darn good performance for very little supply current. It is not rail to rail but you could power it from the +/- 9V of an LT1081 A simple voltage follower and then an inverting amplifier with the noninverting pin at 1V seems like a good way to go. PNP circuit: If you don't mind some drift and can afford to lose gain, you could use a PNP emitter follower with the emitter resistor being a voltage divider up to the +5V supply. I very bad idea: You could use a flying capacitor switching circuit screaming along at 10 or more MHz. > > Now My first idea was to make a JFET input, tried something in spice like this: > ftp://panteltje.com/pub/fet.gif > > There is a lot of spread in individual jfets, and the only negative voltage > available is a tap of the -8 V from a MAX232. The LT1081 is a little more robust than the MAX232. > > So that leaves opamps, and the ones I found (5V rail to rail) > have either low bandwidth, bad slew rates, > incredible distortion, or all of those at the same time. > > So I am looking for a decent quad 5V to 12 supply opamp with low current draw, > only needs to drive a PIC input, say 10k in 25 pF, with low distortion, high > gain, slew rate > 10V us, voltage swing to very near the positive supply, > input impedance 00, input cap < 10 pF. > > I looked at LMC6036, some others, but no good. > Any ideas? > So it is for the scope input amp.
From: Jan Panteltje on 13 Apr 2010 08:07 On a sunny day (Mon, 12 Apr 2010 17:11:10 -0800) it happened Robert Baer <robertbaer(a)localnet.com> wrote in <sbudnULSLNOZKl7WnZ2dnUVZ_qSdnZ2d(a)posted.localnet>: > !!MAXIM parts??? > They actually _exist_??? > Gag me with a spoon! How many MAX232 do you want? :-)
From: Jan Panteltje on 13 Apr 2010 08:15 On a sunny day (Mon, 12 Apr 2010 19:16:32 -0700 (PDT)) it happened speff <spehro(a)gmail.com> wrote in <229e36b4-62f4-4f3c-906f-977275e2f9fd(a)i12g2000vbi.googlegroups.com>: >On Apr 12, 7:33�am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: >> Technical question about good low power rail to rail opamp. > ^^^^^^^^^^^^^^^^^^^^ >Please, JP, this is primarily a political discussion group. hehe :-) >> �I published 'scope_pic' >> �http://panteltje.com/panteltje/pic/scope_pic/ >> and am getting questions as 'to where to connect the input etc..'. >> >> The question is how to make a 1 MOhm, few pF, input with a bandwidth of a= >bout 1MHz, >> with as few components as possible, that also creates a +2V DC offset >> (to get the ADC midrange and allow negative signals), > >Are you looking for gain of 1, or do you need other gains too (like a >real 'scope)? Basically 1, but 10 would be OK too for an extra input range. If I set the internal reference of the PIC to 1.024 V (it is 4 x that now), then 64 pixels becomes 1 V on the LCD, or 16 mV per pixel. With a gain of 10 you get 1.6 mV per pixel. <The >GBW of the amplifier will have to be much higher if you want to see a >1mV 1MHz >signal fill the display. The math works like this for me: the fastest sample speed is 500 uS full screen (128 pixels horizontal). So per pixel 3.9 us. The opamp should be able to go from 0 to 1 V (or 0 to 4 V in the current setup) in 3.9 us. So there is plenty of room. >> So that leaves opamps, and the ones I found (5V rail to rail) >> have either low bandwidth, bad slew rates, >> incredible distortion, or all of those at the same time. > >I guess you could use the +/-8V supplies and avoid the need for a R-R >amp? The +5 and the -8 will probably do.
From: Jan Panteltje on 13 Apr 2010 08:27
On a sunny day (Mon, 12 Apr 2010 19:39:21 -0700 (PDT)) it happened MooseFET <kensmith(a)rahul.net> wrote in <ce35884c-01fe-47e8-921d-83da65f069c4(a)12g2000yqi.googlegroups.com>: >On Apr 12, 4:33�am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: >> Technical question about good low power rail to rail opamp. >> �I published 'scope_pic' >> �http://panteltje.com/panteltje/pic/scope_pic/ >> and am getting questions as 'to where to connect the input etc..'. >> >> The question is how to make a 1 MOhm, few pF, input with a bandwidth of a= >bout 1MHz, >> with as few components as possible, that also creates a +2V DC offset >> (to get the ADC midrange and allow negative signals), > >Op-amp idea: > >The LT1352 gives a darn good performance for very little supply >current. > >It is not rail to rail but you could power it from the +/- 9V of an >LT1081 > >A simple voltage follower and then an inverting amplifier with the >noninverting pin at 1V seems like a good way to go. > >PNP circuit: >If you don't mind some drift and can afford to lose gain, you could >use a >PNP emitter follower with the emitter resistor being a voltage divider >up to >the +5V supply. > >I very bad idea: >You could use a flying capacitor switching circuit screaming along at >10 or more >MHz. > > > >> >> Now My first idea was to make a JFET input, tried something in spice like= > this: >> �ftp://panteltje.com/pub/fet.gif >> >> There is a lot of spread in individual jfets, and the only negative volta= >ge >> available is a tap of the -8 V from a MAX232. > >The LT1081 is a little more robust than the MAX232. Ok, got it, but I already have a whole bunch of MAX232. My other idea of an input was this: +5 +5 | | e e PNP b-------- b PNP c | c +5 | |______| -2 to | | | sets +2 V c | R3 DC shift ---- b NPN e--------1 k ---> 0 to 4V | e ---------- b PNP R2 /// | c R1 | | -8 -8 This is two opposite connected emitter followers, the Vbe drops should cancel, The PNP current mirror adds a constant drop arcross R2 to shift the voltage to the positive side. The input impedanve is huge too. But an opamp is simpler. Have not tried this though. JFETS are nice as input. |