Using 5v Standard to generate a 4v Pk to Pk AC signal with out losing precision?

[rwgast_lowlevellogicdesin]

Ok so I have a DMM Check from voltagestandard.com. I would like to calibrate my DMM's AC though. I had an idea and im wondering if its viable or if im going to lose accuracy the accuracy of the standard?

I was thinking about connecting it output to an mega ohm resistor divider, this would be to shave it down to 4v from 5v, then to an opamp(LM-385) in unity gain to buffer the signal then running it through a TI TLE-2426 (http://www.ti.com/lit/ds/symlink/tle2426.pdf) which is a precision rail divider. At this point I would then have a +2,-2 output which I could pulse with a 555.

My v standard 5.0000 volts exactly, and my DMM is a 40,000 count .05% +5 accuracy/resolution meter. Im not worried about the frequency out of the 555 so much but I need it to +2.0000/-2.000 pulsed waves. So will this circuit ruin the standards accuracy or what? Im afraid of noise and/or loading the standard is going to mess things up.

If this isnt a great idea maybe someone else out there has a better way? Or even if it would work its be nice if there was a way to do this with less chips. I know his DMM check plus does 100hz pulsed AC at -/+2.5v. But to keep my meter in its range I need 4v peek to peek so 2.5 is no good. His business is also to make money so his projects are not open source and he even scratches the labels off certain chips. This is no issue but it basically means I have no idea how he implements AC.

[Seekonk]

At work we made a device that had to trip on the same DC and AC mv value. And it had to be cheap, all out test rigs were cheap.  I remembered an old article I had read years ago.   It put a small lamp in a bridge with 3 resistors.  I played around with resistor values, as the bridge had to have a slight voltage out ( 5V in and 200mv out.  For the 5VDC in I used a 7805 regulator the AC was just raw transformer dropped to about 5VAC.

Switch from DC to AC and you got the same mv reading on the meter with an input that varied on the transformer from 105 to 135AC.  Of course this is thermal regulation bases on the change in resistance of the lamp.  The lamp was placed in foam for thermal isolation.  This is the same as many old oscillators that used lamps.  I'm thinking it is easy to measure DC with accuracy, AC is a lot harder.  So, add an amplifier and adjust so you get exactly 4VDC.  Then switch to AC and you will have the same RMS AC voltage.  The short term thermal drift should be minimal.

Like any good politician, I answered the question I wished you had asked.  Someone else may find this useful.  I thought it was a novel solution to a problem. Hey, I gol lots of mercury wetted chopper relays.

[rwgast_lowlevellogicdesin]

What kind of lamps are we talking about? 12v halogen bulbs, lights from autozone etc, I have a feeling the wrong selection of lights could cause a lot of thermal issues? I can deal with it only staying stable for about a minute, long enough for me to trim the meter.

I wouldnt have thought you and I were the only two on this forum who have any ideas to make an acurate DC to AC conversion. Do all the volt nuts only care about DC??? I know it can be done since Doug (From DMM standard) does it on his boards! He told me he would send me calibration data for 10 bucks if I made a reference that piggy backs off of his but Id rather know how to do this and not worry about needing someone with an 8.5 digit meter meter to help me figure out if its calibrated or not.

[Vgkid]

See this...
www.linear.com/docs/7461
you would need an input of .47v, though I wonder if you could just use a buffer and omit the 20/10 K resistors. That way you could use 1.414v as the reference. I wish the eval board was cheaper .

[mrgregs]

If you went the bridge oscillator way, type 327 lamps were often used - do a search for wein bridge oscillator for circuits, both with and without lamps. However the idea has been refined quite a bit since they used lamps, so Vgkid's circuit is probably the most precise.