Let the breadboarding begin...

[nhmaf]

A few suggestions, in case you haven't already planned on them:

1) One of the 'flaws' in the Twinmax design - possibly they've improved the design lately, but as of several yhears ago they hadn't - was  that it used unregulated battery voltage in the device.   Thus, the error of the measurement was directly influenced by the variation in the battery voltage.   Thus, if you had a "weak" battery int he device, it wouldn't provide the same accuracy as a new battery.

2) Use a "Bridge" circuit design  to minimize component variations messing with your "zero".  I would feed the differential signal from the bridge into an op amp with relatively low gain, maybe just 1. You maybe better off doing the filtering of the first op-amp's output if you have to filter stuff, rather than having the filter circuit's components causing a potential imbalance to the differential input.  Then feed that output into another op amp circuit for more gain, etc.

If you are really interested, I've "acquired" a schematic that may be of interest to you.

[weasel01]

A few suggestions, in case you haven't already planned on them:

1) One of the 'flaws' in the Twinmax design - possibly they've improved the design lately, but as of several yhears ago they hadn't - was  that it used unregulated battery voltage in the device.   Thus, the error of the measurement was directly influenced by the variation in the battery voltage.   Thus, if you had a "weak" battery int he device, it wouldn't provide the same accuracy as a new battery.

I asked that earlier and he said it would be vehicle powered (good foresight methinks) If it's a low current design the system will never know it's there.

[weasel01]

A few suggestions, in case you haven't already planned on them:



2) Use a "Bridge" circuit design  to minimize component variations messing with your "zero".  I would feed the differential signal from the bridge into an op amp with relatively low gain, maybe just 1. You maybe better off doing the filtering of the first op-amp's output if you have to filter stuff, rather than having the filter circuit's components causing a potential imbalance to the differential input.  Then feed that output into another op amp circuit for more gain, etc.

If you are really interested, I've "acquired" a schematic that may be of interest to you.

Your saying a cascaded design for high selectivity and low error (noise)

I think this is a good approach too as it will track very fast.

Also I cant speak for Justin but I'd love to see your schematic ....is it your design?

[nhmaf]

Regulation is needed for "good" accuracy - this won't draw much current, but if the input voltage can vary from 11V to 13.5V, and one expects "12V" in the design, depending on the implementation of the design, one could introduce error of up to 5% or more.  But, depending on how accurate it needs to be versus the cost trade-offs...
maybe one person's "lousy" is "good enough for automotive work"!
 

While I have designed things like this, I haven't done one JUST like it yet.  I do happen to have a schematic that is purported to be the same as that found in a commercial unit.   A quick glance at it appears to indicate that it would work, but I have not verified the "provenance".

I won't post it up here in case of possible copyright issues, but I will email a PDF to anyone here who wants it and PMs me with their email address, for purely educational purposes, of course!

[Justin B.]

I'm trying to keep it as simple and cheap as possible while still having something that will provide usable data.  Right now I'm at the sensor, one dual op-amp, 4 resistors, and a pot.  If the zero floats around with changes in supply voltage (it will be running off vehicle battery) I'll toss in a 78LXX regulator.  I'm basically using a reference design (with a couple of tweaks) from the chip manufacturers app notes.

[Justin B.]

I had a little time today to finish assembling a more portable prototype based on what I had breadboarded.  It's probably been 20 years since I've did anything like this and was surprised I had it wired correctly on the first try!  All I was able to do was power it up, zero it, and test the ports with the Mity-Vac.  Probably I'll have time to hook it up to a running bike next weekend.

Pardon my depth of field issues while I'm still getting used to a macro-zoom lens...  

[Justin B.]

And the initial test circuit...

Oops!  I just saw after posting I have the + and - switched on the op-amp inputs...  

[weasel01]

Justin looks good! Have you tested it yet?

[nhmaf]

To help reduce the  noise from the bike ignition affecting the amplifier's power supply rail I'd suggest that you put a 10 uF or bit bigger, 16V electrolytic capacitor where the +12V and GND come onto the board, and put a smaller, ceramic capacitor, say around 100nF (0.1uF) right next to the op amp chip (keep the legs/traces of the capacitor close or directly soldered to the power and GND pins.   You may need to do more filtering types of signal/power filtering, but this should help reduce some of the "high freq hairy-ness" you'll observe on the op amp's outputs and is good practice.   The MC33272 is a pretty good choice for this type of application, too !

You've got to come up with some sort of catchy name for it, too -
like "JEB" (for Justin's Electronic carb Balancer", or maybe "Bowser-1" or...
 

[Justin B.]

I had thought about a small electrolytic to smooth out the DC, we'll see how this goes.  Another dangerous idea would be to take the sensor and mate it to a continuous little LED bar graph display up in the fairing or on handlebars.  I remember Edlebrock markets (or used to) a display unit that gave an air/fuel ration readout...

[Justin B.]

OK, first real-world test!!

I hooked up the "Gizmo" to my '81 R100 and tweaked idle and off-idle to get a pretty good null reading.  My idle had pretty much been right on but off-idle was off so I messed with the throttle cables until I got it where I wanted it.  My part throttle running was noticeably smoother.

I then hooked up my mercury manometer to check and I was within less than an inch Hg so I finished tweaking with the manometer hooked up.

I have determined that my difference is likely due to the "zero" shifting around with the fluctuating supply voltage found in a running vehicle as Mike had foreseen.  I think if I add a small electrolytic, a load resistor, and a regulator to the circuit this should stabilize the supply voltage and allow me to get a very good sync.

This also leads me to my next hair-brained scheme and that being to turn this into a product that stays on the bike at all times.  I envision this "little black box" residing under the seat somewhere and a small "dashboard" unit with an LED bargraph to allow somebody to constantly monitor their state of sync and tweak as desired.  I need to find about a 7 segmented bargraph array with segment 4 being a different color than the rest.

[Bengt_Phorqs]

Want to check this out on my R90/6 in the next week or so?  I just rebuilt the carbs and using the home-built manometer have them close enough for guvmint work but still a bit off.  I'll be coming to pick up my hat anyhow, so I'll be glad to let you experiment with your In-Sync-erator.  

[Justin B.]

We can see how it does, maybe use a separate voltage source as I doubt I'll have the circuit "enhancements" done by then.

[Semper Gumby]

hey Justin,

I once saw a FM tuner that had two LED (same color) for signal tuning.  When the signal centered both LED were full bright.  As one side dropped off it got dimmer.  Same for the other side.  Simple way to do the carb balance.  When both LEDs full bright = carbs in sync.

good luck.

[Bob_Roller]

Saw this item on the Bing Agency website, finally found out a little more about the Synchromate II  .

http://www.synchromate.com/index.php

Kinda sorta looks like what you're trying to build .