LM43601: +15V to -15V converter doesn't work properly

Nelson Niu

Part Number: LM43601

I followed the Ti web design and created LM43601PWPR as follows, I built 2 pcb those can get -2.7V output only. My -15V load is about 0.3A.

Here is my original bom:

Bill Of Materials May 6,2019 17:32:43

Item Quantity Value Pcb Footprint ManufPN Manuf DgkPN Description Price_1K Reference

1 2 4.7uF_50V C0805 C2012X5R1H475K125AB TDK Co. 445-5980-1-ND Cap MLCC 0805 4.7uF 10% 50V X5R 0.2346 C38,C40

2 1 2.2uF_16V C0603 EMK107BB7225KA-T Taiyo Yuden 587-5835-1-ND Cap Cer 2.2uF 16V X7R 0603 0.08898 C39

3 1 .47uF_50V C0603 UMK107B7474KA-TR Taiyo Yuden 587-3170-1-ND Cap Ceramic 0603 0.47uF 10% 50V X7R 0.04413 C41

4 1 220pF_50V C0805 C0805C221J5GACTU KEMET 399-1127-1-ND Cap 220pF/50V0805 C0G 5% Ceramic 0.02718 C42

5 2 22uF_25V C1206 CL31A226KAHNNNE Samsung EM 1276-3047-1-ND CAP CER 22uF 25V 10% X5R 1206 0.19685 C43,C45

6 1 100uF_50V CD83X105 EEE-FP1H101AP Panasonic EC P19726TR-ND Cap Aluminum Electrolytic 100uF 20% 50V SMD 180MOhm 0.27968 C44

7 1 47uH_1.8A DO3316P DO3316P-473MLD Coilcraft |MPN_994-DO3316P-473MLB DO3316P SMT Pwr Ind 47uH 20 % 1.8 A 0.914 L5

8 1 0603_1.00M R0603 RT0603FRE071ML Yageo YAG5944CT-ND Res 1.00M 0603 1% 1/10W 0.0122 R41

9 1 73.2K R0805 CR0805-8W-7322FT Panasonic EC ERJ-6ENF7322V Res 0805 73.2K Thick Chip 1% 1/8W 150V 0.01367 R42

10 1 0603_105K R0603 RC0603FR-07105KL Yageo 311-105KHRCT-ND Res 0603 105K 1/10W 1% 0.00476 R43

11 1 LM43601PWPR HTSSOP16 LM43601PWPR Ti 296-40221-1-ND IC REG BUCK ADJ 1A SYNC 16SSOP 2.03034 U12

According to the chip AN SNVA722B, After I changed the C26 to 0.1uF, R24 to 140K, R26 to 10K, tried grounding the pin6. I found the LM43601 convertor works after apply external -15V to the LM43601 output (use Cn1). Once it start working, if I remove the external -15V, it keeps working. But it stop working if I only apply the +15V to my board (use J3) from beginning.(Note: the CN1 trick doesn’t work for the PCB without above changes)

My board needs designed use 15V as input J3. Before LM43601PWPR works, I have to use CN1 with +/-15V for the power input.

over 4 years ago

0 Katelyn Wiggenhorn over 4 years ago

TI__Genius 11040 points

Hello,

Would you be able to share the schematic? This will make it easier for us to help evaluate your problem.

Thank you,
Katelyn

0 Nelson Niu over 4 years ago in reply to Katelyn Wiggenhorn

Prodigy 90 points

0 Nelson Niu over 4 years ago in reply to Katelyn Wiggenhorn

Prodigy 90 points

Yes, I did with MS word paste with schematic but those pictures lost. Anyway I just replied with two schematics.
Update: today after I add a 50K from RT pin to GND (not -15V), one board work with light load now. But the 2nd board with heavy (0.3A) load still need external -15V make it work.

0 Katelyn Wiggenhorn over 4 years ago in reply to Nelson Niu

TI__Genius 11040 points

Hello,

I am unable to open this file. Can attach this file as an attachment? I apologize for the inconvenience.

Thank you,
Katelyn

0 Katelyn Wiggenhorn over 4 years ago in reply to Katelyn Wiggenhorn

TI__Genius 11040 points

Hello,

I received your schematic images via e-mail. Thank you for sharing.

When your device is not working as expected, can you measure the VCC voltage? This will signify if the device has started up as expected. Second, if the VCC voltage is expected, can you measure the voltage from SW to GND? This will give us an idea of the operation of the device.

Also, as you mentioned I would expect that decreasing the resistance of R24 and R26 could help. This will increase the current through the FB resistor divider and in turn make it less sensitive to noise.

Thank you,
Katelyn

0 Katelyn Wiggenhorn over 4 years ago in reply to Katelyn Wiggenhorn

TI__Genius 11040 points

Hello,

I want to follow up with you. Were you able to measure the VCC and SW voltages?

Thank you,
Katelyn

0 Katelyn Wiggenhorn over 4 years ago in reply to Katelyn Wiggenhorn

TI__Genius 11040 points

Hello,

Due to inactivity, I am going to close this thread. Please feel free to respond if you have any additional questions.

Best Regards,
Katelyn

0 Nelson Niu over 4 years ago in reply to Katelyn Wiggenhorn

Prodigy 90 points

Sorry I didn't get to this thread days ago. Here is my answers:
1> I still have the startup problem.
2>My VCC is good, variable power supply, I varied it from 12~16V no help for the start up. At same time, my other circuit works ok with this same power supply.
3>When -15v fail to start up, I got ~2.7V voltage with slow ramp up/down.
4>If I apply a -13~-15.5V power to my board, the LM43601 start working, then if I remove the -15V supply, the LM43601 can keep the output. I checked the switch waveform is around 500KHz with 50-50 duty.

0 Katelyn Wiggenhorn over 4 years ago in reply to Nelson Niu

TI__Genius 11040 points

Hello,

Thank you for measuring the VCC. This means the internal device is on. Can you share the oscilloscope waveform of the Switch signal and output voltage?

Thank you,
Katelyn

0 Nelson Niu over 4 years ago in reply to Katelyn Wiggenhorn

Prodigy 90 points

I can't paste picture to this post. How can I send the waveform to you?

0 Nelson Niu over 4 years ago in reply to Nelson Niu

Prodigy 90 points

wf0531.zip

I insert a zip file, including following files:

# Archive C:\temp\waveform\wf0531.zip
2019-06-01 06:10       1207254 fail_outpt-2.9Vzoomin1_13.bmp
2019-06-01 06:10       1207254 fail_outpt-2.9Vzoomin2_14.bmp
2019-06-01 06:10       1207254 fail_outpt-2.9V_12.bmp
2019-06-01 06:02       1207254 working505K+-15v_10.bmp
2019-06-01 06:03       1207254 workingOutput-15.4V_11.bmp

0 Nelson Niu over 4 years ago in reply to Nelson Niu

Prodigy 90 points

I have upload the waveform days ago. Any update?

0 Katelyn Wiggenhorn over 4 years ago in reply to Nelson Niu

TI__Genius 11040 points

Hello,

Thank you for sharing these waveforms.

My first question is when you made the modifications, did you also change C27?

Second, in your waveforms the "working output"'s channel 2 is 1.5V (500mV/division). Should this be 5V/division? Should the second channel be the output voltage?

Also, is there any difference in test conditions between the working images and the not working images? - output load current, PCB board, etc?

Thank you,

Katelyn

0 Nelson Niu over 4 years ago in reply to Katelyn Wiggenhorn

Prodigy 90 points

thank your reply. Answer your question:

1 No. C27 never changed.

2 That's 15V, the backup scope probe used caused that wrong display. Yes, if 2nd ch active, that attached the output.

3 No load difference, same PCB, same power supply used.

0 Katelyn Wiggenhorn over 4 years ago in reply to Nelson Niu

TI__Genius 11040 points

Hello,

Thank your or your reply and clarification. Looking at the waveforms, I believe there may be a stability issue.

The C27 or the feed-forward capacitor is based on all ceramic output capacitance. The output capacitor is an electrolytic. I would start by removing the CFF capacitor. With the ESR of the output capacitor most likely this CFF capacitor is not necessary. You can also try adding in a much smaller CFF in the range of 20-40pF.

Could you share the layout of this design as well?

Thank you,

Katelyn

0 Nelson Niu over 4 years ago in reply to Katelyn Wiggenhorn

Prodigy 90 points

thank you. I'll try that. The board gerber file attached.Gerber7930.zip

0 Katelyn Wiggenhorn over 4 years ago in reply to Nelson Niu

TI__Genius 11040 points

Hello,

Thank you for sharing the gerbers. Please see my recommendations below:

The output capacitor should be placed directly next the inductor output to ground.
The feedback trace should be connected near the output capacitor. This should act as a sense line to the output. Also, the feedback trace should be routed away from the switch node to avoid any interference.
I want to verify that there are thermal vias under the device (DAP) to provide thermal relief. Currently, the DAP does not directly connect to a polygon pour for the ground plane. I would recommend expanding the DAP area in order to improve the thermal relief

Please let me know if you have any questions.

Best Regards,

Katelyn

Power management

Power management forum

LM43601: +15V to -15V converter doesn't work properly