$20 LCR ESR Transistor checker project

[Cosmic_Starlight]

also idk if flashing what seems the M firmware which i think its the latest development series would require any hardware modifications to work with a device that i assume was supposed to have a modified chinese k firmware

Before thinking about new m-firmware, you need to make the ttester itself with k-firmware work. What current does your tester consume when powered by a 9 V battery? By pressing and holding the button. Show a photo of your display from the side of the chips.

Power draw with battery in bit turned off is now 0.10-0.12 mA (no atmega or screen in) turned on with atmega in but no screen is about 0.46 mA

Thats the power draw i meassured before i knew i had two shorted transistors

[Yuriy_K]

Thats the power draw i meassured before i knew i had two shorted transistors

I am interested in the current with faultless transistors, installed microcircuit and connected display. When replacing transistors, you need to find out the position of E-B-C and compare it with the diagram.

If the transistors are working properly, press the button and hold it for more than 2 seconds, release the button, the backlight should remain on if the ATmega is working.

[Cosmic_Starlight]

Thats the power draw i meassured before i knew i had two shorted transistors

I am interested in the current with faultless transistors, installed microcircuit and connected display. When replacing transistors, you need to find out the position of E-B-C and compare it with the diagram.

I will have to buy the replacements for the transistors and ics if i can get those too, maybe replace all caps for ceramics too


Cheapest programmer i found is a usbasp orisbisp one (one thats bare blue pcb no case and hast an extension cable that kind of looks like the one you found on a 3.5" floppy drive and has a other tiny pcb that connects to the other end and has a 6 pin connector on the other side)

Would that work for flashing the atmega with whatever software us required under windows ? (Could do Linux too if i must but am not too used to it)

[snapper]

Since the ay-at has no isp,
you need the ATMEGA8 Minimal System Board in
addition to the usbasp for programming the atm328 !

https://www.ebay.com/itm/192209428607

[Yuriy_K]

Would that work for flashing the atmega with whatever software us required under windows ?

Here is my Windows ISP programmer. For my DIY projects I use a pin header or an adapter with short wires as shown on the right.

[Feliciano]

it seems for the price of another new unit+shipping i could locally get:
another atmega328p (dunno if original or not, nor do i know if that makes a difference)
a cheap programmer for it
and a ST7735B (which is the cheapest one i could find that sort of looks like the original)

It's difficult to find an original ATmega in SMD format nowadays. It's more likely you get a valid one if you buy the DIP version.
The benefit of getting the real one is to be able to flash the original firmware for it (K or M, depending on the device and your preferences). Of course you would need a programming cable and add/construct the interface.

in half the time the new one would take to arrive, (that is if the new one isnt dead too)

If you have somebody near the border and later how to forward it to you, I could sell you one of my generic or already modified units.

[Cosmic_Starlight]

Since the ay-at has no isp,
you need the ATMEGA8 Minimal System Board in
addition to the usbasp for programming the atm328 !

https://www.ebay.com/itm/192209428607

Really ? I cant getthose down here and importing them would be too expensive

Would that work for flashing the atmega with whatever software us required under windows ?

Here is my Windows ISP programmer. For my DIY projects I use a pin header or an adapter with short wires as shown on the right.

That exact usb programmer is the one am looking to get so it should work right ?

it seems for the price of another new unit+shipping i could locally get:
another atmega328p (dunno if original or not, nor do i know if that makes a difference)
a cheap programmer for it
and a ST7735B (which is the cheapest one i could find that sort of looks like the original)

It's difficult to find an original ATmega in SMD format nowadays. It's more likely you get a valid one if you buy the DIP version.
The benefit of getting the real one is to be able to flash the original firmware for it (K or M, depending on the device and your preferences). Of course you would need a programming cable and add/construct the interface.

in half the time the new one would take to arrive, (that is if the new one isnt dead too)

If you have somebody near the border and later how to forward it to you, I could sell you one of my generic or already modified units.

Unfortunatedly i don't have anyone near the border

I saw a video of a spanish guy adding the connector to the ay/at board directly to program it with that usb one so i guess i could to that too ?

(Ive attached screenshots of that video)

[snapper]

Yes, Really! There is no ISP connector on the AY-AT board. You either take the ATM328 out of the socket and put it into the adapter board I suggested for programming (which is connected to the USBASP), or you solder this unsightly wire fiddling to the socket of the AY-AT board every time you program. But do as you see fit!
he also delivers this adapter to Colombia  https://www.ebay.com/itm/404390299187

[Cosmic_Starlight]

Yes, Really! There is no ISP connector on the AY-AT board. You either take the ATM328 out of the socket and put it into the adapter board I suggested for programming (which is connected to the USBASP), or you solder this unsightly wire fiddling to the socket of the AY-AT board every time you program. But do as you see fit!
he also delivers this adapter to Colombia  https://www.ebay.com/itm/404390299187

Am sorry i didnt mean to sound ungrateful, but like i said i might or might not also need to buy another screen so i need to take the screen into the budget of the repair

That device you suggested for its price+shipping i could get just the usbasp programmer locally

And yes i agree its very unsightly but i need to get this thing working first atleast.
i can always spend the time later and make it prettier

I also would like to avoid having to remove it several times in case i flash the wrong fw or didnt set up screen correctly and risk breaking off a leg of it

[hapless]

If you're still thinking about replacing all those yellow capacitors, I really see no reason to do that. You have ceramic capacitors, most are multi-layer, and they don't need to be a different type than they are. If they short out or don't have the right capacitance, that's different, but if they are good, leave them in.

What I would like to see is if you have an LED in the test socket, does it blink a few times several seconds after you apply power like you did before. This is basically just to see if the ATMega is programmed at all.

[Cosmic_Starlight]

If you're still thinking about replacing all those yellow capacitors, I really see no reason to do that. You have ceramic capacitors, most are multi-layer, and they don't need to be a different type than they are. If they short out or don't have the right capacitance, that's different, but if they are good, leave them in.

What I would like to see is if you have an LED in the test socket, does it blink a few times several seconds after you apply power like you did before. This is basically just to see if the ATMega is programmed at all.

the one time i did sort of manage to get it to turn on (as in press encoder led lights up and screen is solid white) but holding down the power button did nothing i had to remove power to turn it off

what i do need to replace is the two s9012 transistors at the very least, maybe the other one, and maybe the 7550 too just to be sure

i already tried feeding 5v to atmega vcc with screen on and same deal, screen is solid white, i dont think the led thing would work at all since doesnt the device ask you to calibrate it upon first turn on ?

[carrascoso]

Yes, Really! There is no ISP connector on the AY-AT board. You either take the ATM328 out of the socket and put it into the adapter board I suggested for programming (which is connected to the USBASP), or you solder this unsightly wire fiddling to the socket of the AY-AT board every time you program. But do as you see fit!
he also delivers this adapter to Colombia  https://www.ebay.com/itm/404390299187

The author of this video is me.   
The cables with the connector soldered FIXED directly to the board, is super comfortable for programming.
Not having to disassemble the screen, remove the AtMega from the socket, plug it into the programmer or into the breadboard to program it.
And after programming the AtMega, put everything back together in the tester.
This way, you can program the AtMega quickly without so much hassle.

[hapless]

[...] doesnt the device ask you to calibrate it upon first turn on ?

I think it will try to test first and complain about being uncalibrated later. Since I always blow out the pre-programmed firmware as soon as possible, I have no memory of how it works. But I don't think it's any different than what the k firmware does.

The programming header can be installed directly onto the board or affixed nearby and doesn't really need to be as complicated as in the video, but keep wires from the board to the header short. Permanent stray capacitance and resistance are not something you want to add a lot of.

[Cosmic_Starlight]

[...] doesnt the device ask you to calibrate it upon first turn on ?

I think it will try to test first and complain about being uncalibrated later. Since I always blow out the pre-programmed firmware as soon as possible, I have no memory of how it works. But I don't think it's any different than what the k firmware does.

The programming header can be installed directly onto the board or affixed nearby and doesn't really need to be as complicated as in the video, but keep wires from the board to the header short. Permanent stray capacitance and resistance are not something you want to add a lot of.

i will try and go downtown this week and buy replacements for all transistors and the 7550 and TL431  (closest replacement ive been able to find for the 7550 locally online is a LM2931AZ-5.0RAG whose pinout is different so i might have to do some fiddling like bending the leads or soldering some jumper wires and then shrink wrap them)

also buy all caps just in case one is shorted or bad, buy 3 crystal oscillators while am it too (8-16-20)

and most importantly another spare atmega and the usbisp programmer for it and the connectors and wires to solder the female port to the tester

also does anyone have a list lying around of the list of capacitor values this thing is supposed to use ? i suppose i could also look at mine note them all down and just say in the store give me x ammount of capacitors that say 102 on them and such
but i think it would be best if i just straight up asked for the specific values (that the chinese AY/AT clone is supposed to have from factory atleast)

also i saw a video of a guy that had a similar issue to mine (turns on red led is on but his screen is solid blue, tho his screen pcb is blue not green like mine) and he said in the comments he later fixed it by replacing the crystal that he was supplied with the wrong one

i cant do the led test right now since both S9012 are shorted between E and C

ive attached a photo that i was dm'ed on reddit from another guy would there be a problem if i did it similarly like this using super thin wiring ? (closest i have is some old cat5e wires that i used to use on a protoboard)

[Yuriy_K]

also buy all caps just in case one is shorted or bad, buy 3 crystal oscillators while am it too (8-16-20)

i cant do the led test right now since both S9012 are shorted between E and C

I did not understand why there are two S9012 pnp transistors in the circuit and one? The other two S9014 npn - this is clear from your circuit. The author of the k-firmware writes that the highest accuracy of his firmware during measurements is guaranteed by an 8 MHz quartz. And this is true, but I managed to make some changes to the source texts and increase the accuracy for a 16 MHz quartz (see the examples of my firmware on page 277). But this is the limit for which you still need to select some elements of your circuit. As an example, 22 pF capacitors in a quartz generator should be reduced to 15-18 pF, etc. The example of connecting the programming connector you showed will not allow you to properly calibrate and work normally with the tester. It needs to be unsoldered. Here is an example of my firmware on a display similar to yours.

[Cosmic_Starlight]

also buy all caps just in case one is shorted or bad, buy 3 crystal oscillators while am it too (8-16-20)

i cant do the led test right now since both S9012 are shorted between E and C

I did not understand why there are two S9012 pnp transistors in the circuit and one? The other two S9014 npn - this is clear from your circuit. The author of the k-firmware writes that the highest accuracy of his firmware during measurements is guaranteed by an 8 MHz quartz. And this is true, but I managed to make some changes to the source texts and increase the accuracy for a 16 MHz quartz (see the examples of my firmware on page 277). But this is the limit for which you still need to select some elements of your circuit. As an example, 22 pF capacitors in a quartz generator should be reduced to 15-18 pF, etc. The example of connecting the programming connector you showed will not allow you to properly calibrate and work normally with the tester. It needs to be unsoldered. Here is an example of my firmware on a display similar to yours.

Wait a minute so you mean to tell me that its not as simple as just swapping the crystal for a faster one and recompiling the fw but i would also need to change capacitors values as and other things as well ?

Also could you please tell me what screen is that (what specific model)

only ones ive been able to find locally online are the ST7735B that has 11 pins so is no good to me and another ST7735S that has 8 pins so it will fit but idk if the pins are in the right order

[Feliciano]

For 8MHz -> 16MHz you change the crystal and build/flash the firmware. 20MHz is not stable enough, so we don't recommend it.
The 22pF Capacitors are there for helping the oscillation. Maybe we could fine-tune that, but normally it's not required.

And, like Yuriy_k said, the AY-AT clones I've seen uses two NPN and one PNP (being the other TO-92 the regulator). Maybe that was the error in the first place? You could trace the voltage path to confirm whether the polarity would it be right for one type or the other.

[Cosmic_Starlight]

also buy all caps just in case one is shorted or bad, buy 3 crystal oscillators while am it too (8-16-20)

i cant do the led test right now since both S9012 are shorted between E and C

I did not understand why there are two S9012 pnp transistors in the circuit and one? The other two S9014 npn - this is clear from your circuit. The author of the k-firmware writes that the highest accuracy of his firmware during measurements is guaranteed by an 8 MHz quartz. And this is true, but I managed to make some changes to the source texts and increase the accuracy for a 16 MHz quartz (see the examples of my firmware on page 277). But this is the limit for which you still need to select some elements of your circuit. As an example, 22 pF capacitors in a quartz generator should be reduced to 15-18 pF, etc. The example of connecting the programming connector you showed will not allow you to properly calibrate and work normally with the tester. It needs to be unsoldered. Here is an example of my firmware on a display similar to yours.

For 8MHz -> 16MHz you change the crystal and build/flash the firmware. 20MHz is not stable enough, so we don't recommend it.
The 22pF Capacitors are there for helping the oscillation. Maybe we could fine-tune that, but normally it's not required.

And, like Yuriy_k said, the AY-AT clones I've seen uses two NPN and one PNP (being the other TO-92 the regulator). Maybe that was the error in the first place? You could trace the voltage path to confirm whether the polarity would it be right for one type or the other.

i am a dumbass i got confused because the numbers were very close youre both absolutely right, there are two  NPN S9014 and one PNP S9012, and the shorted ones between E and C are both the S9012 T3 and T1 S9014

and so a ST7735 screen in 128x160 should be fully compatible with this firmware (provided i get the version that has 8 pins not 11) ?

[Per Hansson]

I bought himself DTU-1701. I painted a complete schematic diagram
https://yadi.sk/d/yW8xa5NJgUo5z/BSideESR02(DTU-1701)/Schematic%20diagram
Red color means changes made to the scheme:
1. Replace the frail  7150 on the LP2950-5.0(better MCP1702-5.0)
2. Replace  TL431 on LM4040A (R2A)
3. Added ceramics capacitors 10mkF on LP2950 input  and Vcc processor(pins 4-6)
4. Control divider resistors R21-R22 for voltage replaced 47kOm, control point moved to the collector Q2 (optional).
5. Capacitor C10 deleted - normal work without it.
6. Replace quartz at 16MHz (optional).
Revision board 3.1 - changes in comparison with earlier BSideESR02 virtually none except for a correct wiring ICSP connections-now all your contacts conveniently separated by the connector. Photo board to upgrade and after it will be lower.

Firmware:
New English and Russian k-firmware revision 813 for this clone  https://yadi.sk/d/yW8xa5NJgUo5z/BSideESR02(DTU-1701)/Firmware/Karl-Heinz%20K%C3%BCbbeler%20

Hello Indman, I have a question about point 4 in the above quote.
If I've understood it right moving the R21-R22 divider to Q2 would get rid of the 9µA quiescent current that the Bside ESR02 draws when it is off right?
But what does changing the resistors from 1M to 47k improve?
Also the moving of the battery negative wire to the power input jack what does that improve, is it for better routing?

[CaptainBucko]

ive attached a photo that i was dm'ed on reddit from another guy would there be a problem if i did it similarly like this using super thin wiring ? (closest i have is some old cat5e wires that i used to use on a protoboard)

That was me! You could definitely use CAT5e wires, it just needs to be thin enough to reliably solder to each pin. The connector should be glued down so that the wires do not suffer metal fatigue and break off. You can use use any connector you like but you may have to make up an adapter cable from the ISP adapter. For this purpose I prefer solid core wire, as it bends and holds it shape. I use wire wrap wire myself.

The ISP adapter I use is this one: https://www.ebay.com.au/itm/353547444898

[Cosmic_Starlight]

ive attached a photo that i was dm'ed on reddit from another guy would there be a problem if i did it similarly like this using super thin wiring ? (closest i have is some old cat5e wires that i used to use on a protoboard)

That was me! You could definitely use CAT5e wires, it just needs to be thin enough to reliably solder to each pin. The connector should be glued down so that the wires do not suffer metal fatigue and break off. You can use use any connector you like but you may have to make up an adapter cable from the ISP adapter. For this purpose I prefer solid core wire, as it bends and holds it shape. I use wire wrap wire myself.

The ISP adapter I use is this one: https://www.ebay.com.au/itm/353547444898

lmao small world huh... and thats the exact usb programmer am eyeing on getting, if i cant get a port to solder to the atmega i will have to use my old protoboard and flash it that way

tho i was adviced against soldering wires to the atmega since they could throw off the calibration or make it have issues under normal working conditions

next thing you know the chinese seller i bought this from might chime in here too lmao since you and the spanish guy in took the yt screenshot from has also chimed in

[indman]

Hello Indman, I have a question about point 4 in the above quote.
If I've understood it right moving the R21-R22 divider to Q2 would get rid of the 9µA quiescent current that the Bside ESR02 draws when it is off right?

Yes, you are correct in your understanding of this decision.

But what does changing the resistors from 1M to 47k improve?

In fact, this replacement will not improve anything, so you may not use it. But from an engineer's point of view, the output resistance of the power supply we are going to measure must be matched to the input resistance of the Atmega's measurement port. It is for this reason that the author chose a 10k/3.3k divider to supply 9V and also to reduce the maximum allowable input voltage below 5V.

Also the moving of the battery negative wire to the power input jack what does that improve, is it for better routing?

This solution allows you to simply hard disconnect the 9V battery if you connect external 9-12V power. You don't have to do that if you want to.

[Per Hansson]

Thank you Indman for the explanations, I really appreciate it!
Sometimes I wonder how my own brain works though, reading the description of the file: Hiland-M644-Booster-Mod.png
Under the repository: https://github.com/madires/Transistortester-Warehouse/tree/master/Hardware
It's wonderful when you Google something and find an explanation: written by yourself!
Edit: here is a photo of the way I did it, just removed R21 and replaced it with a through hole resistor, so no need to cut the trace this way...
Works great and now 9µA quiescent current is down to only 0.2µA

[Cosmic_Starlight]

so i finally went downtown today to find replacement parts for it and here its what i could get

S9014 ----> BC547
HT7550 -----> LM78L05
S9012 ----> N3906

only as a last ditch effort i tried one last store if they had any of the original references, and i bought the last two S9012 they had in stock, think i bought two TL431 too jut in case, they didnt have any S9014 nor 7550

bought all new caps, and two crystals of 8 and 16mhz

bought a spare atmega too but all i was told is that they were generic, another store had it for twice the price think generic too but it came with a bootloader already i just bought the cheaper one

no store had a programmer for it, the one store that had one like it was red pcb with a huge zif socket and its price was more than a new gm328 unit so i didnt buy it obviously

am fully aware all the replacements i bought have a different pinout so i also bought some 1.5mm heatshrink tubing

i just realized i might have made a mistake and asked for a replacement of a 7750 not a 7550, i dont know if the replacement i bought for it will work or not, or if i should just go and buy a different one again that has a lower dropout voltage (hopefully i didnt apply too much heat to the 7550 when desolderingg it)

[CaptainBucko]

I don't see a 6pin J3 connector here, i.e. I have to make do with the Chinese firmware which is pre-installed, right ?

As mentioned by others, you can convert your PCB from a LGT8F328 to Atmega328p to run the latest v1.53m firmware, with a bit of extra wiring to add an ISP header to the PCB.

Your PCB is slightly different from mine, but these photos should give you some guidance. Tnx to others before me who created solutions this is based on.