Interesting G-120 amp design...

Hey folks... Someone forwarded this site to me. It's got details on a home-built scanner amp design that can drive the G-120's at 24K and still give solid results. Pictures on the site tell the story - they look great!

Don't know how expensive they'd be to build, but my guess is that it would be cheaper to build these things vs trying to buy a pair of Turbo Track amps on E-bay. (Even if the Turbo Track amps can go up to 30K, these still look pretty good at 24K.)

Anyone else ever see these? (Well, besides you, Brian! :p)

Adam

Additionally, his name on pl is pishtak. I emailed him not too long ago and he was super nice, sent me the complete schematics, complete board layout in .pdf format as well as .brd and .sch.

I remember him saying the parts were only about $50. As far as the boards go, that would be a different story. Small quantities from a manufacturer would be spendy and every time I have tried a double sided board, I have jacked it up (not to mention all the through board jumpers). I am ready to give this a try though. Anyone have any tips on double sided boards or moreover, a cheap manufacturer that would not charge an arm and a leg and first born for 10 pieces?

There might be more to this, Buffo. I seriously doubt that he could get 8 degrees at 24K from an LM675T driven by only 20 volt rails. This is the kind of thing that I would have to see with my own two eyes...

On the page he mentions using Showtacle software. I don't know that software very well, but I have seen in their advertisements something like "fixed 35Hz scan rate" as an attempt to try to prevent damage to scanners. Since I don't know Showtacle software, it makes me wonder if the scan rate can really be adjusted separately. It could be that this guy thinks he is scanning at 24K but really isn't...

Anyway, if he will be present at a future LaserFreak meeting, that might be a time when I can see this with my own eyes, and also plug in our softare and do some additional testing...

Bill

Leave it to Bill to pop high speed scanning bubbles with his good intentions to make sure we get what we pay for!!:p:D

Glad you mentioned something about it, Bill. Was actually thinking about picking up a set of G-120's from Aron Bacs at FLEM just to play with this. But if you don't think it's a viable option, then I'll hold off. (After all, I already have enough projects on the back burner as it is... I don't need one that is destined to fail!)

Adam

I am still going to try this. It won't be that expensive and it will give me a new project. I'll see what kind of speeds I can get. I'll mainly be using Spaghetti and LaserBoy since I am poor. I can send the rig to Bill for testing when completed.

Cool Brian... Keep us informed of your progress!

Adam

That is a fairly advanced g120 amp. It looks like a hybrid of a g660 by GSI, as far as the shutdown stuff is concerned, and a early cambridge.

G660 was GSi's high end amp before minisax sumdiff etc, and was a standard for 15-20 years. What it lacks in speed it will make up in precision as unlike most web designs, it does use both sides of the position sensor.

I'd look at subbing a LM3886 for the output stage for a bit more power and bit less heat, plus you could route the shutdown signal in to the lm3886s mute pin.

Steve

Steve, for an LM3886 substitute, what would you recommend for supply voltage in this type of application?

Quote:

---

Originally Posted by absolom7691

Steve, for an LM3886 substitute, what would you recommend for supply voltage in this type of application?

---

The 3886 can handle plus minus 35 V

Steve

Quote:

---

Originally Posted by buffo

Glad you mentioned something about it, Bill. Was actually thinking about picking up a set of G-120's from Aron Bacs at FLEM just to play with this. But if you don't think it's a viable option, then I'll hold off.

---

I just don't think that this thing will really work "as advertised".

It may be possible to tune it up to 24K for very small angle... BUT, then what will happen when you put some content in with large angles? Experience tells me that it will massively overshoot and possibly (probably) break a torsion rod.

Then, it may be possible to tune this amp up to work at larger angles (which is what the G-120 is supposed to do), but then small signal response will be crap (like 15K or below, as he shows in some of his other pictures).

This amp has no "angle intelligence", so it can either be tuned up for small signals, or large signals, but not both at the same time. This amp will also "sing", depending on the mirror/mount you put on it.

So, in the end, I think (gulp) Chinese scanners and amps would work better than G-120s and this amp, and probably be less expensive too.

But in any event, if someone has some money burning a hole in their pocket, sure, go ahead and get it. It never hurts to experiment and sometimes that's the best way you learn.

Best regards,

William Benner

Maybe I will heed your warning for now Bill. Besides, I may be getting some Catweazle II's soon. Anyone have any opinions on those? These would be the 30K catweazles. Based on the tests I saw on the Pangolin site, they looked somewhat comparable to the g-120s running 30K on a Turbo Track.

Quote:

---

Originally Posted by Pangolin

This amp has no "angle intelligence", so it can either be tuned up for small signals, or large signals, but not both at the same time.

---

This is interesting. How does one implement "angle intelligence" in a scanner amp? Is it some sort of amplitude filter that kicks in at a given level and changes the tuning for larger steps? Or maybe an integrator that continually adjusts some portion of the amp in response to the magnitude of the signal change? I'm really curious about this...

Quote:

---

This amp will also "sing", depending on the mirror/mount you put on it.

---

I'm not sure what you mean by that, but it sounds bad! Is this some kind of resonance that will develop in the rotor of the scanner?

Quote:

---

So, in the end, I think (gulp) Chinese scanners and amps would work better than G-120s and this amp, and probably be less expensive too.

---

Wow... *That* is really saying something! I've already decided not to pursure this, but I am still interested in learning more about scanner amps in general.

Adam

I think one question, for me, still remains. What is it about the Turbo Track that makes it so fast?

Could one build an amp that employs the same acceleration feedback that the 124 open loop amp has and combine it with an efficient feedback circuit?

Quote:

---

Originally Posted by absolom7691

I think one question, for me, still remains. What is it about the Turbo Track that makes it so fast?

---

"Benner Magic" :D:D

Seriously though, Bill put a lot of thought into that design, and it's just one hell of a well-built amp. But that's why they were expensive when they came out, and why people are reluctant to sell them now. They still perform really well, so people hang on to them. (Think Mercedes vs Yugo)

As for the specific technical details behind the amp's continued success, you'd have to ask Bill. :o Scanner amp design theory is *way* over my head...

Adam

Very interesting amp. Hmm, schematic diagram is very close to Chan's amateur amps (the first link on that page), even components values are the same.
I'm also about making new amps for my galvos. I saw on many drivers two power amp ICs. I suppose that is bridged current feedback drivers. Does anyone know how to do that?

Sergey, LM3886 is just a big inverting opamp and with split supplies, can be treated as such.
Thats why many drivers have dual lm3886

A little knowledge is dangerous, test with a speaker or wire wround resistor before you go galvo:

but here:

http://apex.cirrus.com/en/pubs/appNote/Apex_AN20U_D.pdf

if you google " lm3886 bridged with dc servo", you might find your goal.


Also national semi conductor ap note bpa200

Steve

sTEVE

Still I love OPA549 series, except the price :D
Piotr.K

Okay, let me add a few words to the main subject; to my opinion the best way to go with fast & universal galvo drive is DSP. There's no fixed parameter schema for a galvo driver to be both fast and accurate in small & large step area. It takes non-linear functions, so preferably parameters should be fetched from previously filled (self-learned, if implemented) table of parameters for each step required.
And it's commercially available ! - x-drive by MediaLas, perhaps others. Too complicated for DIY.
Piotr.K

Quote:

---

Originally Posted by buffo

This is interesting. How does one implement "angle intelligence" in a scanner amp? Is it some sort of amplitude filter that kicks in at a given level and changes the tuning for larger steps? Or maybe an integrator that continually adjusts some portion of the amp in response to the magnitude of the signal change? I'm really curious about this...

---

the amp in question has the second thing you mention.. if anything the angle intelligence probably amounts to a properly placed capacitor or op-amp-capacitor pair in the feed-forward section.

Quote:

---

I'm not sure what you mean by that, but it sounds bad! Is this some kind of resonance that will develop in the rotor of the scanner?

---

I am reminded of the notch filter that had to be added to some earlier galvo amps to compensate for a resonance in the feedback pickup of the galvo. If the G120s have a strong 'Q' self-resonant frequency then a similar notch filter might be appropriate.

Quote:

---

Wow... *That* is really saying something! I've already decided not to pursure this, but I am still interested in learning more about scanner amps in general.

Adam

---

Given that the author has scope plots proving operation at 15pps at what he claims as 40 degrees (no reason to assume he's lying) then there's reason to expect he would have noticed if his software wasn't actually drawing at 24kpps.

The amp documented on his page may not be perfect, but it's still pretty good, considering the design is released free. Definitely worth considering building and tweaking if there were a glut of amp-less galvos available.

Mixedgas:
"LM3886 is just a big inverting opamp"

LM3886 is just a big opamp. Inverting or non-inverting option depends on external components. At this moment my drivers use them as non-inverting power opamps.

"test with a speaker or wire wround resistor before you go galvo"

I prefer high-power resistors as active load. Speaker costs much more than resistor.

"but here:..."

Wow! Special cases of the bridge connection... It seems that what I need. Thank you Steve!

LesioQ:
"to my opinion the best way to go with fast & universal galvo drive is DSP"

Compare cost of signal processor (or FPGA as that) and cost of three chips with 12 opamps. Most of all I dont need the highest precision and the fastest speed. I only want to accelerate my current galvo drivers.


Sergey

[QUOTE=Sergey;68069]Mixedgas:
"LM3886 is just a big inverting opamp"

LM3886 is just a big opamp. Inverting or non-inverting option depends on external components. At this moment my drivers use them as non-inverting power opamps.

"test with a speaker or wire wround resistor before you go galvo"

I prefer high-power resistors as active load. Speaker costs much more than resistor.

A speaker at my local electronics place is 5-8$ for a decent power level. Wire wound ceramic "fireproof" resistor is 3$ for 2-3 watts, speaker, in terms of mass/per financial unit, is better deal.

Strange , but true.

I too, would use the resistive load.

Steve

So, all that being said, the question comes back up. Is this a decent amp? Bill doesn't seem to like it and I am inclined to listen to the man becuase of his vast experience, but then again, even Einstien had to revise a few of his theories after they were published. I need something though because I am barely getting 15K out of my G120s at 8 degrees and I can't seem to find a Turbo Track either. Unless someone has a couple of Minisax amps they want to part with, this is my next option. I am hoping for 20K at 10 to 15 degrees. If this amp can't do that, or may damage my galvos' torsion bars, then I don't want to attempt it.

I had almost half of it's large step time with MiniSAX. MiniSAXes have bridged output, though they lack hf-damping (current dev.)
I'll count my MiniSaxes over weekend and check if have too many.

This design appears OK, nothing suprising.... ok, except it's claimed speed :D
Piotr.K

As I can see from that drivers schematic diagram HF-damping is not used(R10 and C39). In my drivers which are 9K max I cant get any effect from HF-damping. Is it limiting by galvos frequency range?

HF damping helps to reduce the oscillative nature of galvo response during breaking after large jump. If Your acceleration and deceleration peak current is not high enough this may be true that HF won't do any visible good. Are You suffering from the dip after fast slope command like square wave ? This becoms visible when tuned for highest speed only. If not - Youur amp is probably not able to push G120 really hard, so HF-damping wont be of use. Are You sure You're running G120 coils in parallel, not in series (as LaserFX suggest as I remember ?).
I've heared that there's also a limit of G120 rotor magnetic saturation ...
Piotr.K

Quote:

---

Originally Posted by drlava

Definitely worth considering building and tweaking if there were a glut of amp-less galvos available.

---

Well, I don't know if I'd call it a glut, but Aron Bacs sells refurbished G-120's for around $100 a pop. At the last FLEM he brought a box of about 50 or so!

He's going to be at FLEM on Saturday; maybe I'll plant a bug in his ear to check out this thread. Might be a few folks interested in picking up a pair or two.

Adam

Magnetic saturation? I heard Bill speak of this. Is this a permanant effect due to too strong a field or is it the theoratical limit of the magnetic properties that can be exerted on a G120 rotor?

Quote:

---

Originally Posted by absolom7691

Magnetic saturation? I heard Bill speak of this. Is this a permanant effect due to too strong a field or is it the theoratical limit of the magnetic properties that can be exerted on a G120 rotor?

---

That would be a short-term effect, I think. It would manifest as " no more rotational force with increased magnetizing coil drive current".
Piotr.K

Quote:

---

Originally Posted by LesioQ

Are You sure You're running G120 coils in parallel, not in series

---

Well, my galvos are not G-120. They are home-made, but principles of work are the same, as I guess.
Probably I cant achieve the highest acceleration because of my PSU - It cant supply more than 2-3A RMS to each driver(5-6A total). And galvos coils have low active resistance - about 0.5 Ohms for each galvo.

It's quite old thread now, but found it few days ago.
Please allow few words about my "baby":

Quote:

---

Originally Posted by Pangolin

There might be more to this, Buffo. I seriously doubt that he could get 8 degrees at 24K from an LM675T driven by only 20 volt rails. This is the kind of thing that I would have to see with my own two eyes...

---

It will be pleasure for me if you can look on it :-)

Quote:

---

Originally Posted by Pangolin

On the page he mentions using Showtacle software. I don't know that software very well, but I have seen in their advertisements something like "fixed 35Hz scan rate" as an attempt to try to prevent damage to scanners. Since I don't know Showtacle software, it makes me wonder if the scan rate can really be adjusted separately. It could be that this guy thinks he is scanning at 24K but really isn't...

---

You are right about Showtacle software, but for testing I'm using Chan's controler mentioned on my web. This controler outputs point at specified PPS speed without any
restrictions. Anyway gento from laserfreak built this amp and tested it with
pangolin sw/hw. here are the results and these are even better:-)

Quote:

---

Originally Posted by Pangolin

Anyway, if he will be present at a future LaserFreak meeting, that might be a time when I can see this with my own eyes, and also plug in our softare and do some additional testing...
Bill

---

Hopefully there will be chance ... Maybe this year ?

Quote:

---

Originally Posted by Pangolin

I just don't think that this thing will really work "as advertised".

---

This AMP is far from perfection of TT, but it can scan ILDA test pattern
at specified speed and angle as advertised here and here and here.
Believe or not, it was proven here and it is enough for me as satisfaction/feedback.

Quote:

---

Originally Posted by Pangolin

It may be possible to tune it up to 24K for very small angle... BUT, then what will happen when you put some content in with large angles? Experience tells me that it will massively overshoot and possibly (probably) break a torsion rod.

---

Today I tested it with smaller mirrors and mirror mounts and it can scan 12K ILDA Test
Pattern at 18deg optical at 24Kpps. Scanning angle was adjusted to 18deg optical with input gain. That means input signal with amplitude +/-10V will drive it to max. 18deg optical. After tuning, quadrature square wave can be scanned without overshoot. If the input gain is then adjusted to 40deg optical there will be overshoot that's clear because it was not tuned to this angle. But normally one will not do this. If yes, there is still protection (over position shutdown) which will protect the galvo and it just works great. On laserfreak I mentioned that once I forget to assemble C16 responsible for damping. Servo gain was almost at maximum. Without overposition shutdown torsion bar would have been destroyed in a fraction of a second. But it wasn't.

Quote:

---

Originally Posted by Pangolin

Then, it may be possible to tune this amp up to work at larger angles (which is what the G-120 is supposed to do), but then small signal response will be crap (like 15K or below, as he shows in some of his other pictures).

---

I asked about max. speed of G120D at 40deg optical with TT here , but it was not answered by you. Did you tested maximum speed of G120D with TT at 40deg optical ?

Quote:

---

Originally Posted by Pangolin

This amp has no "angle intelligence", so it can either be tuned up for small signals, or large signals, but not both at the same time. This amp will also "sing", depending on the mirror/mount you put on it.

---

I don't know what you mean with angle inteligence. Maybe short explanation from your side ?

Quote:

---

Originally Posted by Pangolin

So, in the end, I think (gulp) Chinese scanners and amps would work better than G-120s and this amp, and probably be less expensive too.

But in any event, if someone has some money burning a hole in their pocket, sure, go ahead and get it. It never hurts to experiment and sometimes that's the best way you learn.

Best regards,

William Benner

---

My AMP was designed in 2005. At that time no chinese scanners were available at prices which are valid today. My AMP was designed from almost ZERO knowledge about scanning/lasershow. There is still things to be improved :-)

My Amp has also built in pre-emphasis. Maybe it is not on the correct place in
schematics. But it's there and only thanks to this pre-emphasis it can work at specified
speeds.

If I will have the chance to get again pair of G120D for 100USD,
I will redesign and implement pre-emphasis at correct place.
Also there will be bridged output stage implemented (thanks to mixedgas).
http://www.cirrus.com/en/pubs/appNote/Apex_AN05U_E.pdf
Bill was also very nice when giving this explanation.

Quote:

---

Originally Posted by LesioQ

HF damping helps to reduce the oscillative nature of galvo response during breaking after large jump. If Your acceleration and deceleration peak current is not high enough this may be true that HF won't do any visible good. Are You suffering from the dip after fast slope command like square wave ? This becoms visible when tuned for highest speed only. If not - Youur amp is probably not able to push G120 really hard, so HF-damping wont be of use. Are You sure You're running G120 coils in parallel, not in series (as LaserFX suggest as I remember ?).
I've heared that there's also a limit of G120 rotor magnetic saturation ...
Piotr.K

---

If you look into http://www.jrbtechnical.com/pdf/t2manual.pdf you will find out,
that HF damping is not being used for G120D. I think it's because G120D has torsion bar and CT series doesn't have. About thermal limitations of G120D look here.

Best Regards

Stefan