Brand new 300W inverter can't support 100w load.

Thread Starter

Davizuco

Joined Jul 12, 2018
27
Hi all,

I bought a 300W 12/200 modified sine wave inverter for my van, but it starts beeping and goes off when trying to manage about 100W. I opened a dispute, and seller sends me a working unit (not the same board), but still want to repair the first one.

I tryed testing it with a 60W light and everithing looks ok, but if I switch on another one, The inverter starts beeping, alarm led switches on, and both lights go off.

IMG_20200520_161846.jpg

I don't have much experience, but taked a look arround and everithing looks ok... checked mosfet and they look ok too, no capacitors conducing, etc..

Finally, I decided to make some inverse engineering on 12v stage. It has KA7500B controller managing two mosfet and in charge of triggering the alarm. The error is triggered by a small circuit, left-hand side of the buzzer, two zener diodes a npn transistor and some other simple components, connected to pin 1 of KA7500B, it is, if I'm not wrong, an error comparator pin (pins 1-2 and 15-16 in the KA7500B)
Alarm Circuit.jpg
When alarm triggers, N13 transistor starts conducing and I can see about 4v at it's Emitter pin (B).
Measuring transistor's base with a cheap oscilloscope, switching on a 220V 60W ligth I can see a lot of noise, and when the second test light was switched on and the inverter goes off the noise stops as shown in the image:

Disparo Fallo entre ZD5 y N13.PNG
Transformer also warms up, may it be making noise and that noise triggers the transistor?

Thanks in advance.
 

Alec_t

Joined Sep 17, 2013
14,280
One possible cause of premature cut-out could be if the current shunt, used to sense load current, has a resistance higher than it should be, perhaps because of poor joints at its ends. Worth having a look-see? I'm guessing the voltage across that shunt is monitored by a comparator somewhere.
 

Thread Starter

Davizuco

Joined Jul 12, 2018
27
Hi Alec,

Checked all resistors with my multimeter and all measures are OK.

The only "stange" thing I found is in the other error circuit, wired to KA7500B's 16 pin, It doesn't look to be related with the problem sice it's transistor didn't conduce during the tests, but with the inverter switched on, no load, I can see about 0,7V behind ZD1 zener diode (Diagram point C), It's supposed to be 0V. Could ZD1 be defective?

Alarm Circuit pin 16.jpg
Anyway, as I said, It doesn't seem to do anything with the problem since no changes had seen on pin 16.
 

Dodgydave

Joined Jun 22, 2012
11,284
Classic design chip, that uses pins 1,2 as the voltage regulator and pins 15,16 as the current sense,.

You can measure the reference voltage on pin 15 , and as the Sense voltage rises on pin 16, when under load. The current can be altered by adjusting the reference voltage on pin 15 .

As previously described, the current shunt could be higher in resistance, this can be cured by resoldering.

Here is a typical circuit for a 30A 24V .

IMG_20200521_110749_9~2.jpg
 

Alec_t

Joined Sep 17, 2013
14,280
Checked all resistors with my multimeter and all measures are OK.
I'd be surprised if your multimeter can accurately measure resistance values of only a few milliOhms. What value did you measure for the current shunt resistance?
If you measured resistances in circuit they won't be accurate values, because there are parallel conducting paths.
 

Thread Starter

Davizuco

Joined Jul 12, 2018
27
I'd be surprised if your multimeter can accurately measure resistance values of only a few milliOhms. What value did you measure for the current shunt resistance?
If you measured resistances in circuit they won't be accurate values, because there are parallel conducting paths.
This is why I missunderstood (and this shouldn´t surprise you ) and weren't measure the right ones. I was thinking "what shunt? I can't see anyone", I'm just learning.. sorry.
You are talking about two R003 resistors I found right now underneath a pile of white silicone.

As you said it´s a tiny resistor to be measured.. this afternoon I will resolder it and see what happens.

Thank you !! ;)
 

Thread Starter

Davizuco

Joined Jul 12, 2018
27
Shunt resistors resoldered, now, connecting a 120W load, the inverter beeps, 220v output goes off, but 2 or 3 seconds later it stops beeping and starts working as expected. It is the same behavior every single time I change from a 60W load to 120W.

I tryed to resolder Shunts again, despite it looks ok from first time, and why not, all components connected to 1 and 2 pin. Then, tested it again, but there was no difference.

Is it normal? I mean, it's not a high quality unit.. may be this is all it can afford? The good one also beeps (for half a second, and 200V doesn't goes off) sometimes when a high load is switched on.
 
Last edited:

Alec_t

Joined Sep 17, 2013
14,280
It looks as though the shunt isn't the culprit then (unless a wrong value shunt has been installed). Perhaps Chinese Amps are smaller than usual :).
 

gerty

Joined Aug 30, 2007
1,305
Are you using a cigarette lighter plug to power this up? I've seen high resistance connections in the power plug that causes a voltage drop when a load is applied. Metering the primary voltage as the load changes should tell you if that's a problem,
 

Thread Starter

Davizuco

Joined Jul 12, 2018
27
Are you using a cigarette lighter plug to power this up? I've seen high resistance connections in the power plug that causes a voltage drop when a load is applied. Metering the primary voltage as the load changes should tell you if that's a problem,
Really??? :D

At the begining I was testing it connected by a screwed-up cable to the second battery in my van, but last days I was using a portable battery with a cigarette lighter plug, as @gerty says.... It' working now!!!!

Thank you all!!!

I think resoldering did the job, but faultly using cigarette lighter plug maked me think it didn't.
Another noob failure I won't forget ;)
 

Thread Starter

Davizuco

Joined Jul 12, 2018
27
I'm going to put here information that someone with a similar investor might find useful.

On normal working state (no error):
  • KA7500b Pin 2: 0.061 volts
  • KA7500b Pin 15: 5 volts (Vref pin 14)

12 Volts Stage, full circuit:

IMG_20200522_170339.jpg
 
You may be seeing the initial current surge of the light bulbs triggering the circuit. A cold bulb may draw several times the current when cold. You could add a capacitor across the monitoring circuit to slow the response.
 

Thread Starter

Davizuco

Joined Jul 12, 2018
27
There is a free place on the board marked as a 100uF 35V capacitor to filter 12V power source at zeners circuit side. I was tempted to add it, but I dont have it now.
 
Top