-
Well fuck that.
Been writing a design description I inherited all day. Figured I had a decent understanding but when dotting the 'i's my calculations were always off compared to the simulations. Dived deeper and finetuned my calculations. Rather to my dissappointment and puzzlement, the calculations moved away from the simulations rather than towards it.
So, finally, my eye caught this...
My brain always registered this as an intrumentation amplifier with a factor of 5. Only now I've found it misses the 'normal' resistor (100k in this case) to ground from the inverting amplifier. Malvino and TAO tell me nothing. So... I'm not sure how to interpret this. Is this a really clever design I'm not getting or should the other guy be flogged for forgetting the resistor?
Not asking for someone to do my homework, if this circuit is an actual thing, just a keyword I can google on would be awesome.
-
Quote from: m k
A pen plotter, maybe they are visiting north shortly.
Roland DXY 990
https://www.leboncoin.fr/informatique/1956772681.htm
He is never going to visit you, because he is not SELLING it, he is LOOKING for one !
Dang,
I didn't even consider the possibility.
Where it is said?
Well in the body of the ad it says : " recherche plotter Roland DXY-990 "
" Recherche " in French means " Looking for ".
The price of 125 Euros he mentions is probably the kind of price he might be willing to pay for one, I presume...Quote from: m kHP 400H
https://www.leboncoin.fr/collection/2110669646.htm
So the item is reserved to the buyer when seller get the money.
What if buyer is a no show?
Well if the buyer doesn't show up, too bad for him ! He already paid for the item so if he changed his mind at the last minute and doesn't want it anymore, the seller keeps the item of course, and can sell it TWICE, more money in the pocket !!
To be honest I have never used this feature on that website, so I don't know. I guess in this case the buyer can cancel his order and gets his money back.... or not. I really don't know.
-
(...)
TE update: 5316b has shipping label created. Investigating oscillator options. Perhaps should buy a 10-pack of those CTI oscillators..
There are also China surplus "VECTRON C4550A1", some of them even have the V-Ref output option (they sadly all look the same from outside).
Mine have arrived and were bought when they were only slightly more expensive then the CTI, which seems to have changed recently
A quick check indicates that 3 of 10 seem to posess the option. You'll have to decide for yourself if that sample size is representative enough.
This guy made nice A-Dev Graphs for them and gave some more details.
TE Update: 5316b package is with carrier. It's only going more or less across the metro area here, but freight costs were so low that diesel fuel spent picking it up would have cost more. Hope for tomorrow, but realistically Thursday.
Yes, that too is quite interesting. My Russian is a bit rusty, so the link was "mostly for pictures" -- but, it seems that extra insulation improved things, right?
-
https://ampnuts-com.translate.goog/vectron-c4550a1-0213/?_x_tr_sl=ru&_x_tr_tl=de&_x_tr_hl=de&_x_tr_pto=wapp(...)
TE update: 5316b has shipping label created. Investigating oscillator options. Perhaps should buy a 10-pack of those CTI oscillators..
There are also China surplus "VECTRON C4550A1", some of them even have the V-Ref output option (they sadly all look the same from outside).
Mine have arrived and were bought when they were only slightly more expensive then the CTI, which seems to have changed recently
A quick check indicates that 3 of 10 seem to posess the option. You'll have to decide for yourself if that sample size is representative enough.
This guy made nice A-Dev Graphs for them and gave some more details.
TE Update: 5316b package is with carrier. It's only going more or less across the metro area here, but freight costs were so low that diesel fuel spent picking it up would have cost more. Hope for tomorrow, but realistically Thursday.
Yes, that too is quite interesting. My Russian is a bit rusty, so the link was "mostly for pictures" -- but, it seems that extra insulation improved things, right?
I prefer the integrated translate function in the Chrome browser for websites in other languages as it is almost seamless.
For all other browsing Firefox is used.
TL;DR: They appear to be SC cut, insulation improves stability. -
Vref : we have voltaing action going on, success !!
I got lucky and was a simple fix, though I wasted some time pulling the data sheet of every chip on the board (6 of them !) and probe around... and fixing the battery leads that broke off 4 times
Eventually I noticed that hey... am I dreaming or is this push button wiggling when I push on it !
What do you know, factory forgot to solder it all together !! Its 4 pins were bent underneath trhe PCB, which was kinda holding it place... but as far as electrical connection, not so good !
30 seconds and 4 joints later... we have life in this thing !!
Grabbed my most accurace DMM a Metrix MX56C , which is 0.025% (+/- one count IIRC) .
Label at the back of the Vref says they checked it with an Agilent 34401A.
Looked on the web the specs of the thing. On its 10V range which is I guess what they used to get maximum resolution, that's where the meter offers it's best accuracy it looks like : 0.0015%
So a bit better than 10 times better than my DMM. Not stellar a difference, but good enough I guess.
Results :
MX56C / 34401A
2.5V : 2.5015 / 2.5013
5V : 5.002 / 5.0015
7.5V : 7.504 / 7.5029
10V : 10.004 / 10.0026
Wow, quite good hey ? Very close to what the sticker says... so maybe they are genuine measurments the factory made, not just some random sticker/values...
If it is, then looks my MX56C is quite good ??
Looks good to me anyway... maybe I could be even closer to the sticker values, if the temperature were the same as they used at the factory (again, assuming e believe the sticker...).
Sticker says 25°C, but it's only 19°C in the house, according to the thermostat built-into the heater which is 4 meters away from the bench.
As Tggzzz sad, would be cool to log the data over a longer period of time to observe the drift. Don't have the capability to automate this right now, lab is not that sophisticated just yet.. a week ago it was still completely unusable... give it more time please !
-
Just received a 100x scope probe, NOS, reasonable money.
Worthless unless you have the frequency derating chart.
It's an Elditest GE 3121, 150MHz, 2kV, though I don't want to ever be within a mile of anything oscillating at that frequency and magnitude! Fabrique en Allemagne, so decent quality I should think.[/color=teal]
Worthless how? I already said I won't be going anywhere near frequencies like that at higher voltages. The highest it'll see is the inside of a smps. In any case the datasheet will be available from the manufacturer I'm sure; Germans are fastidious about such things. -
As Tggzzz sad, would be cool to log the data over a longer period of time to observe the drift. Don't have the capability to automate this right now, lab is not that sophisticated just yet.. a week ago it was still completely unusable... give it more time please !
Amy data is better than none. Just leave the reference and meter on, and note the time/voltage whenever convenient.
That won't catch noise (thermal or popcorn) but it will show drift. -
As Tggzzz sad, would be cool to log the data over a longer period of time to observe the drift. Don't have the capability to automate this right now, lab is not that sophisticated just yet.. a week ago it was still completely unusable... give it more time please !
Amy data is better than none. Just leave the reference and meter on, and note the time/voltage whenever convenient.
That won't catch noise (thermal or popcorn) but it will show drift.
... but... I am not sure the Vref will stay on for say 24 hours, with an LED lit very bright, with a small battery ?
Doesn't this thing turn itself off after a while, to save battery ??
Will try I guess but...
-
Do download it and examine the frequency derating and you'll see what I mean. They rolloff big time at just a few MHz which of course you won't see in the std SMPS.Just received a 100x scope probe, NOS, reasonable money.
Worthless unless you have the frequency derating chart.
It's an Elditest GE 3121, 150MHz, 2kV, though I don't want to ever be within a mile of anything oscillating at that frequency and magnitude! Fabrique en Allemagne, so decent quality I should think.[/color=teal]
Worthless how? I already said I won't be going anywhere near frequencies like that at higher voltages. The highest it'll see is the inside of a smps. In any case the datasheet will be available from the manufacturer I'm sure; Germans are fastidious about such things.
But good on ya for getting a 100x HV probe and playing it safe.
-
As Tggzzz sad, would be cool to log the data over a longer period of time to observe the drift. Don't have the capability to automate this right now, lab is not that sophisticated just yet.. a week ago it was still completely unusable... give it more time please !
Amy data is better than none. Just leave the reference and meter on, and note the time/voltage whenever convenient.
That won't catch noise (thermal or popcorn) but it will show drift.
... but... I am not sure the Vref will stay on for say 24 hours, with an LED lit very bright, with a small battery ?
Doesn't this thing turns itself off after a while, to save battery ??
Will try I guess but...
What prevents you from connecting a laboratory power supply instead of the battery? -
As Tggzzz sad, would be cool to log the data over a longer period of time to observe the drift. Don't have the capability to automate this right now, lab is not that sophisticated just yet.. a week ago it was still completely unusable... give it more time please !
Amy data is better than none. Just leave the reference and meter on, and note the time/voltage whenever convenient.
That won't catch noise (thermal or popcorn) but it will show drift.
... but... I am not sure the Vref will stay on for say 24 hours, with an LED lit very bright, with a small battery ?
Doesn't this thing turns itself off after a while, to save battery ??
Will try I guess but...
What prevents you from connecting a laboratory power supply instead of the battery?
Or just permanently charge the battery. -
ON THE BENCH TONIGHT: Temp PSU for 32in Cinema Display
Today's been a helluva day; after losing a hour sleep to DST (invented by some crusty old bureaucrat who cut one end off his blanket and sewed it to the other end to make the blanket longer...
), I put in a solid 10 hours hoeing out garage and shed to make garage usable space and to move machinery/shift gears from snow-removal mode to lawn-care mode. Took only enough time off to eat and pop handfuls of Ibuprofens... 
Anyhoo... so after it's all done, I sit down to relax with dinner and check in here and... that damned Cinema display is fucking up again, only this time the PSU brick is doing the tick-tick failure mode. FML...
So I bring the thing down to the dwagon-cave and pop it apart again; but for the life of me, I can't figure out what holds the aluminum heatsink on the other side to the inner plate all the semis are attached to. This means that even if I did find some questionable caps, I'd have no way of getting them in/out to replace them.
I decided that rather than go all Samsonite Gorilla on the damn thing getting it apart, maybe the better idea would be to try a different PSU altogether, just to see if the problem was in the PSU or the display.
Sadly, the big Lite-On brick I was thinking of was 20V/11A out rather than 22V/10A like I thought, and it simply would not power up connected to the Cinema Display which brick is rated 24.5V/6.4A, (I suspected too high surge current... so not promising) so I had to drop back & punt.
I did toy with the idea of modding one of those 750W HP server PSUs so I could make a 24V stack... but I've done it before and I know that doing it even marginally safely is a lot more assache than I was up for after the day I'd had. Also, knowing these PSUs, if there is something wrong in the display itself causing the tick-tick fail, it's not going to shut down on short protection, it's just gonna keep pumping current up to ~120% of rated capacity, which in this case is 80-ish amps.
So... I decided to just make the connector from the Apple power brick into a pigtail, and put that on the 24V/15A PSU from one of my 3DPrinters. It took about 5 minutes to get it out, so already a couple hours ahead. A quick rummage in my quadcopter building stuffs yielded some nice 12ga silicone wires the perfect length; a bit overkill, but as this is a temporary fix, I'll always be able to reclaim them for their original purpose. I definitely don't need to worry aboot voltage drop... 
Oh, it feels sooo goood having my MetCal back; need to tin 12ga wires and all you can lay paws on is a 1.2mm chisel tip...? No problem...
Soldering those fat wires down is just as easy; a dab of rosin and BAMMM! smoky joy.
Took a moment to blow the PSU out (did I mention I am actually able to get to my air compressor in the garage now...?
); it has been powering a 3D Printer for 2 years, so it's past due. Noticed some primo Chinglish on the warranty seal...
Not a finished product by far; this is quick & dirty just to git 'er done. Bench test showed it has no problem powering up with the Cinema Display connected; so I adjusted the output to 24.5V as per the rating of the original PSU brick. Just to make Zucca and mansaxel happy, I did at least immobilize both cords with a quick zip-tie so no wire-wiggle-induced bare conductors peeking out from under the terminal strip.
But NO! I did not terminate anything; these screw terminals are designed for bare wires, DAMMITT! Worse, just a few wraps of green tape to keep the angry pixies going where they're aspoda go. THE HORROR!!!
These LED Lighting PSUs tend to be a bit noisy; even ones like this with thermal fan control. So for the sake of being able to hear if there's any tick-tick going on inside the Cinema Display, it had to go under the desk.
For now, the display appears to power up without issue; no funny noises or smells, just that lovely 32-in near-IPS screen.
We'll see how it holds up, and maybe shop a generic power brick if it does.
Or if I get to wanting that 3D printer back online...
mnem
I am gonna be sore in the mornin'...
-
Well fuck that.
1) Only a negative feedback --> circuit is linear
2) if it is linear you can apply the superimposition of effects
3) Two inputs: one through R7, let's call it "Vin". The other one is through R8 and it is called "2V5" (whatever it means)
4) One output = Opamp output, let's call it "Vout".
The trick of point 2) is to ground one input and calculate the output. Do the same with the other input and then sum the two effects to get the real output.
Vout_DC = -5 * "2V5" + 6 * Vin
In high f you get about
Vout_hf = - "2V5"/200 + Vin
Not sure why R7 is there.... Also do not understand the "clever design" -
ON THE BENCH TONIGHT: Temp PSU for 32in Cinema Display
How you could resist to not dremel the Liteon PSU open?
Ops you opened it... well done!
Nive work, even with Z-tie :-) -
Last I looked, working with this display still $800-1000. Specialty shops asking considerably more. It's because of the software it can run if you have the licenses.Yeah, I kindof feel the same... my only valid assessment of comparative value is the $$$ of the thing as configured when new...
I don't think I'd be game to look at the gear I have in that light. IMO, it's only worth what its worth on today's values.
Yeah, I certainly wouldn't pay that for it, but I've never been their primary customer base.
mnem
-
I taught myself CP/M by poking around a DEC unitized workstation with two 5¼" floppies. I taught myself MS/DOS with the blue and red books, mostly by reading and following where the pointers in the books led; I only had access to a PC at the local library. I completely skipped over the first versions of Windows; I couldn't see where they were any better than BOB.if I ever got my Kung-Fu up to snuff.
I wish you stop underrate yourself here. You've got the mind for it, I can tell.
The problem is that you're likely to be frustrated by the myriad things-that-look-they-might-work in most Linuxen. The complete lack of oversight, taste and regulation in the Linux world makes things a bit chaotic there. Like visiting a cliché oriental bazaar and trying to do a trend analysis of the carpet market. Every subsystem has at least 17 slightly varying graphic config utilities that each consume 8x the compute resources that the underlying actual daemon/whatever will, just to look cute. And none of them work. As soon as one gets into tinkering mode (and Dragons do!), the only path is text editor on config file.
You instead probably need to install something that is very simple and actually will deliver to promise, like one of the BSDen.
Somewhere in there I did in fact mess around with RedHat LINUX; almost got proficient with that and later Ubuntu on a original PPC White MacBook. ALMOST.
It's just that... well... I've run out of patience for having to open the hood and twirl wrenches every few days. Same is true of my vehicles as well; which is why when I left Houston, I left Franken-Cruiser™ behind.
mnem
I think I've done enough penance for being born with The Knack. -
The Lite-On brick is industrial-duty... power for a line of Kiosk PCs I had a service contract on aboot 5 years ago. It actually screws together. I have little doubt, since the 3DP PSU seems to run the Cinema Display just fine, that the display was drawing more current trying to make up for those missing 5 volts, and that was why the Lite-On brick kept going into shutdown. It'll run 3 LiPo fast-chargers at once without breaking a sweat.ON THE BENCH TONIGHT: Temp PSU for 32in Cinema Display
How you could resist to not dremel the Liteon PSU open?
Ops you opened it... well done!
Nive work, even with Z-tie :-)
mnem
*toddles off to ded*
-
Do download it and examine the frequency derating and you'll see what I mean. They rolloff big time at just a few MHz which of course you won't see in the std SMPS.Just received a 100x scope probe, NOS, reasonable money.
Worthless unless you have the frequency derating chart.
It's an Elditest GE 3121, 150MHz, 2kV, though I don't want to ever be within a mile of anything oscillating at that frequency and magnitude! Fabrique en Allemagne, so decent quality I should think.[/color=teal]
Worthless how? I already said I won't be going anywhere near frequencies like that at higher voltages. The highest it'll see is the inside of a smps. In any case the datasheet will be available from the manufacturer I'm sure; Germans are fastidious about such things.
But good on ya for getting a 100x HV probe and playing it safe.
OEM appears to be French. Roll-off starts about 300 kHz. Down to about 300 V at 8 MHz. Manual here:
https://bkpmedia.s3.amazonaws.com/downloads/manuals/en-us/GE3121_manual.pdf
Current model @ £79 from Farnell https://uk.farnell.com/elditest/ge-3121/probe-osciloscp-2kv-150mhz-x100/dp/1532989
Note that the BNC is not insulated. The manual says: "Probe assemblies must not be used for measurements on mains circuits."
It's IEC1010 CAT I which is not for use on ciruits conneted to the mains. I've always thought this misleading because most people think "CAT" raings are for mains use.
This probe is not suitable for use on the primary circuits of mains SMPSUs -
Nice catch.
Do download it and examine the frequency derating and you'll see what I mean. They rolloff big time at just a few MHz which of course you won't see in the std SMPS.Just received a 100x scope probe, NOS, reasonable money.
Worthless unless you have the frequency derating chart.
It's an Elditest GE 3121, 150MHz, 2kV, though I don't want to ever be within a mile of anything oscillating at that frequency and magnitude! Fabrique en Allemagne, so decent quality I should think.[/color=teal]
Worthless how? I already said I won't be going anywhere near frequencies like that at higher voltages. The highest it'll see is the inside of a smps. In any case the datasheet will be available from the manufacturer I'm sure; Germans are fastidious about such things.
But good on ya for getting a 100x HV probe and playing it safe.
OEM appears to be French. Roll-off starts about 300 kHz. Down to about 300 V at 8 MHz. Manual here:
https://bkpmedia.s3.amazonaws.com/downloads/manuals/en-us/GE3121_manual.pdf
Current model @ £79 from Farnell https://uk.farnell.com/elditest/ge-3121/probe-osciloscp-2kv-150mhz-x100/dp/1532989
Note that the BNC is not insulated. The manual says: "Probe assemblies must not be used for measurements on mains circuits."
It's IEC1010 CAT I which is not for use on ciruits conneted to the mains. I've always thought this misleading because most people think "CAT" raings are for mains use.
This probe is not suitable for use on the primary circuits of mains SMPSUs
Had me look up the ones we carry and they are CAT II again.
http://www.pintek.com.tw/productDetail/land-ctop-2/index/pscsn/17073/psn/19272 -
Marco Reps is selling some of his stuff:
https://www.ebay-kleinanzeigen.de/s-bestandsliste.html?userId=15084259
NAWTS
Not the custom 8.5 digit CERN multimeter (I would buy that
), and the Keithley 155... though it does make you wonder what he is collecting funds for.
-
As Tggzzz sad, would be cool to log the data over a longer period of time to observe the drift. Don't have the capability to automate this right now, lab is not that sophisticated just yet.. a week ago it was still completely unusable... give it more time please !
Amy data is better than none. Just leave the reference and meter on, and note the time/voltage whenever convenient.
That won't catch noise (thermal or popcorn) but it will show drift.
... but... I am not sure the Vref will stay on for say 24 hours, with an LED lit very bright, with a small battery ?
Doesn't this thing turns itself off after a while, to save battery ??
Will try I guess but...
What prevents you from connecting a laboratory power supply instead of the battery?
Because you are TOO QUICK BU, because you are too quick !
When you asked that last night I clearly was not even sure I was going to conduct this experiment or not. If I did then I did not even know if it would time out /power off by itself or not.... so yo are asking me why I didn't not try to correct a problem that 1) I was not even sure I would go even be concerned with and 2) I didn't even know even existed or not !
Gimme time !
But let's say I went to do what you suggest, the problem is that I would not know WHERE to connect thelab supply exactly ! As I said I don't have a schematic for the thing, nor a board layout, so I don't know how the thing works... where to feed power to get the thing to work and operate properly (get the push button to cycle thourgh the voltages to begin with !) while at the same time bypassing the "auto-off" feature, especially since the two functions might well be implemented in the same chip ! So it's either all or nothing, no cherry picking allowed !Quote from: TggzzzOr just permanently charge the battery.
That's assuming the device is designed to turn the auto-off feature off, when the charger is plugged in. I don't know that.
Anyway, that was to reply last nights comments.... now after a good night of sleep I have some good news :
I did decided to conduct the experiment, and turned the Vref on before going to bed. Just at midnight.
Now I just woke up. At 9h00, the Vref is STILL up and running, so no auto-off features it looks like !
It's my DMM that turned itself off...
So I have now two data points, 9 hours apart.
How much drift ? Zero.
I set it up to 2.5V because a) according to the datasheet of the AD584K that's the output that gives the best performance and b) it's the only output that allow me to use the full 5 digit resolution of the DMM, making it is easier/more likely to catch drift.
So, at power up got 2,5015 (2.5013 on the sticker).
9 hours later overnight, DMM still reads 2.5015....
-
Nice catch.
Do download it and examine the frequency derating and you'll see what I mean. They rolloff big time at just a few MHz which of course you won't see in the std SMPS.Just received a 100x scope probe, NOS, reasonable money.
Worthless unless you have the frequency derating chart.
It's an Elditest GE 3121, 150MHz, 2kV, though I don't want to ever be within a mile of anything oscillating at that frequency and magnitude! Fabrique en Allemagne, so decent quality I should think.[/color=teal]
Worthless how? I already said I won't be going anywhere near frequencies like that at higher voltages. The highest it'll see is the inside of a smps. In any case the datasheet will be available from the manufacturer I'm sure; Germans are fastidious about such things.
But good on ya for getting a 100x HV probe and playing it safe.
OEM appears to be French. Roll-off starts about 300 kHz. Down to about 300 V at 8 MHz. Manual here:
https://bkpmedia.s3.amazonaws.com/downloads/manuals/en-us/GE3121_manual.pdf
Current model @ £79 from Farnell https://uk.farnell.com/elditest/ge-3121/probe-osciloscp-2kv-150mhz-x100/dp/1532989
Note that the BNC is not insulated. The manual says: "Probe assemblies must not be used for measurements on mains circuits."
It's IEC1010 CAT I which is not for use on ciruits conneted to the mains. I've always thought this misleading because most people think "CAT" raings are for mains use.
This probe is not suitable for use on the primary circuits of mains SMPSUs
Had me look up the ones we carry and they are CAT II again.
http://www.pintek.com.tw/productDetail/land-ctop-2/index/pscsn/17073/psn/19272
Not sure I understand... your website says CAT 2 only for DC, but CAT 1 for AC. Mains is AC of course, so CAT 1...
No ?
Sorry if I misread you or Rob
-
It's just that... well... I've run out of patience for having to open the hood and twirl wrenches every few days. Same is true of my vehicles as well; which is why when I left Houston, I left Franken-Cruiser™ behind.
mnem
I think I've done enough penance for being born with The Knack.
Perhaps that is so. What I'm thinking is that once one gets into running such things on a daily basis, sortakinda "critical mass", it becomes second nature. And you indicate you're not there. (A lot of "second nature" of course us internalizing the chores as "natural". Which may or may not be a Good Thing... ) -
Also do not understand the "clever design"
Neither do I but you helped me quite a bit. A lot better than a rubber duck, much appreciated. Now calculated = simulated
-
Quote from: Tggzzz
Or just permanently charge the battery.
That's assuming the device is designed to turn the auto-off feature off, when the charger is plugged in. I don't know that.
Anyway, that was to reply last nights comments.... now after a good night of sleep I have some good news :
I did decided to conduct the experiment, and turned the Vref on before going to bed. Just at midnight.
Now I just woke up. At 9h00, the Vref is STILL up and running, so no auto-off features it looks like !
It's my DMM that turned itself off...
So I have now two data points, 9 hours apart.
How much drift ? Zero.
I set it up to 2.5V because a) according to the datasheet of the AD584K that's the output that gives the best performance and b) it's the only output that allow me to use the full 5 digit resolution of the DMM, making it is easier/more likely to catch drift.
So, at power up got 2,5015 (2.5013 on the sticker).
9 hours later overnight, DMM still reads 2.5015....
Spec, for the AD584 alone, is "25ppm/1000hours". Presumably they mean the first 1000hours, since usually the drift is assumed to reduce over time - hence the usual sqrt(1000hours) LM399 specification.
So, you should keep it powered for a few months and/or get a better well-aged meter.