INN2215K Application Problem

Dear jazzclassics,

Because the INN2215K SR MOSFET driver typically uses a supply rail of 4.4 V, a MOSFET with too high a threshold voltage, VGS(th), is not suitable.  It is recommended that the SR MOSFET have a threshold voltage well below the drive level.  See Figure 1 for preferred MOSFET characteristic.

So, please try an AO4294 FET instead of the IRF540N for Q1 and see if performance improves.

-PI-ODO

dear PI-ODO,

thank you very much for the reply with graph.

Do you think it possible to increase the output power of the inn2215  to a little more than 30 W? This will allow the adapter to be suitable for many more PD applications.

I did read in the datasheet "A MOSFET with 18 mW RDS(ON) is a good choice for designs rated for 5 V, 2 A output. The SR MOSFET driver uses the secondary SECOND￾ARY BYPASS pin for its supply rail and this voltage is typically 4.4 V. A MOSFET with too high a threshold voltage is therefore not suitable and MOSFETs with a low threshold voltage of 1.5 V to 2.5 V are ideal although MOSFETs with a threshold voltage (absolute maximum) as high as 4 V may be used." The IRF540 has an abs max Vgs of 4 V. I thought it might be ok for this circuit. Should it work if I just use a diode and remove the SR mosfet?

My build so far is like the picture attached.

Thanks again for your reply. I think PI is doing one of the best products for the battery charger manufacturers.

Dear jazzclassics,

Thank you for your comments.  To support 30W (and greater), we plan to release the next-generation InnoSwitch family later this year. 

Although operation with just a diode instead of the SR FET is possible, it would be a shame not to reap the efficiency benefits by using a suitable SR FET.

-PI-ODO

Hi PI-ODO, 

Thank you very much for your reply.This is the first project I do with a  SMPS. Therefore, your support is highly appreciated. 

My demo board now is up and running, but it still has some problems. No matter what the AC input voltage is, the output voltage is five volt, and it is stable, which is good. However, the output current can be no more than 1.8A. When it's over this limit, The switching will be stopped. It can be resumed by reducing the load.What's more, When output 9V, switching will be stopped at 0.6A. And the circuit can not output a 12 volt, It responds  to the interface IC request by swinging up the voltge to 12, and then quickly in miniseconds goes into protection, which can not be recovered until you recycle the AC input. And it does not matter if it's with a load at the output or not.I have tried and replaced the zener diode(ZD1 in the data sheet) at the bias circuit to a 30volt zener diode, which did increase the current up to 2.3 A at 5V. I read， from the data sheet, that this zener diode is used to perform the OVP，but it does not explain how it work in detail. Could you explain more in details how it works and why it has to be a 20volt zener?

I have the similiar problem. I can't get 12 V output in this circuit. Also I can't achieve work as QC3.0 charger.

there are very scarce literature about the INN2215. Here in China, a similar model INN1227 is more common to see in the consumer adapters, XIAOMI for example used widely the INN1227 in its MI5 and MI5 SP phone chargers. There is even no INN1227 data sheet on the PI Web site.

Hope the PI experts can stand out and explain in more details why the components should have those values, and how it would fail if some component is out of  the range.

Hi jazzclassics,

The primary sensed OVP function utilizes a Zener diode from the rectified and filtered bias winding voltage supply to the PRIMARY BYPASS pin.  Selecting the Zener diode voltage to be 6V above the bias winding voltage (28 V for 22 V bias winding) gives good OVP performance for most designs, but can be adjusted to compensate for variations in leakage inductance. The resistor in series with the OVP Zener diode limits the maximum current into the BYPASS pin. 

When the current into the PRIMARY BYPASS pin exceeds ~7.6mA, the part will shut down, thus implementing the OVP function when the Zener voltage is exceeded.

Hopefully, this clarifies the operation of this circuit.

-PI-ODO

Hi TALLmen,

Could you please describe in detail exactly what is happening ? And could you please send me your exact schematic.

PI-AP

Hi Jazzclassics,

On page 3 the section on 'PRIMARY BYPASS Pin Output Overvoltage Latching Function' and on Page 9 the 'Overvoltage Protection' does explain exactly the same thing which was explained in the previous post about the use of Zener diode and how it can lead to the latch off.

PI-AP

Hi PI-ODO,

thank you for clarifying this. I did not realize that this bias circuit plays both output OVP and bias OVP until read your reply.  It truly helps. 

Now it remains to me a question why that OVP actually acts as a output current limiter. I achieve 5V 2.2A output with a 30v zener, but merely 1.6A with a 24V zener. And it says in the datasheet that for 22v bias, the zener should be 28v. however no actual implementation use higher than 24v. I see your official design examples use 20v. So, what is the optimal value for the bias voltage, what is the best voltage to supply the BP pin?

Another question, in the datasheet T1 is suggested to be a PQ20 transformer.  but your design examples all use RM8.Can a PQ2016 be used in this circuit? Then, what would the output power be? Can you post an example spreadsheet with PQ2016 acting the T1.

Hi jazzclassics,

For a good design example, please refer to DER494 at https://ac-dc.power.com/design-support/reference-designs/design-examples/der-494-195-w-quick-charge-20-30-compliant/.

It may be easier to start from there (with a design fully validated to be compliant) rather than wrestle with so many component alternates.

-PI-ODO

Dear PI-AP,

We solved our problems. DER-494 is quite useful and it was helped us, after we bought this design board.

But still there is one question. How can we check that our charging equipment is working with QC3.0. Are there some tests?

Also maybe you'll advise me where can we make the sertification of this charging that it complies the requirments QC3.0?

Dear TALLmen,

Power Integrations has tools for testing QC 3.0 compatibility. You might want to contact the nearest sales office to provide a tool for testing. 

UL (Underwriters Laboratories) provides QC certification and testing services. You should contact the nearest UL Lab for certification.

PI-AP

I managed to get a INN1227 and replaced the INN2215 and replaced the transformer with one from another vendor. The circuit is running much better now, but still needs some tuning. The major problem is that the output power drops sharply as the output voltage rises from 5.31V to 12.6V. Here is the measurement result.

I probed the drain and found the overshoot can be as high as 180V, this is too high compared with the design exaple DER494, which is merely around 60V. Meanwhile, the output ripple also far from satisfactory. I attach the waveform here. The yellow line is the drain and the cyan line is the output ripple measured with a 100MHz DSO.

Can somebody help to let me know what is wrong and how to improve here? Thanks a lot.

Test items

5V 0A Test

5V 2.3A Test

5V 3.5A Test

12V 0A test

12V 1A test(auto shutdown when increasing the load to let output go over 1A)

Hi Jazzclassics,

The high peak drain voltage and ripple depend on the components selected and the transformer design details. Could you please share yor schematic and the transformer deisgn spreadsheet and construction?

PI-AP

Hi PI-AP,

Thanks for your reply. The schematic is attached.The schematic is almost the same as the one in the INN2215k datasheet, except for that the SR fet is IRF540N. The transformer is build by following transformer construction instructions in the DER494, except that all the primary windings are soldered on the bobbin pins 1-3 and 10-12.

Is the high drain spike a problem with the snubber circuit？ The problem is not only the high drain spike, but also the limited current at high voltage output. It should output at least 1.5A, but could make barely 0.95A. What frequency should the switch work for a 12V out? Thanks a lot.

Jack

Hi PI-AP,

Could you kindly take some time to answer my last questions? I didn't use any transformer design spread sheet, I just follow the instructions in the DER494. TCould you do a quick experiment with the IRF540 as the SRfet and let me know if it cause a problem?

By the way,have you released any product/design example that supports the PD3.0.Thank you.

Hi Jazzclassics,

We have DEE533 in our website for USBPD applications. 

What is the power supply doing once the output current is not going beyond a point? Is it doing Auto restart or latch off or no response? The drain waveforms in your attached images show negative regions of operation with 10 V per division, are they captured correctly in DC mode of the oscilloscope?

Your present design is getting power limited. The max power would determine the input capacitor values. Usually we follow 1:2 ratio. Using the transformer design spreadsheet helps to know if the design is optimum for delivering the desired amount of power and it flags warnings if anything is violated or not met. The spreadsheet also tells the frequency of operation for delivering the right amount of power. So I would recommend you use the PI-Expert tool in our website for transformer design.

If the leakage of the transformer is high, then it could lead to high spike, so please let me know how much leakage vs the primary inductance value do you have in your transformer. If the snubber circuit is not optimized then it could also lead to high spike too. The type of diode used (standard, fast, ultra fast) in the snubber circuit also affects the peak drain voltage, standard diode giving the lowest peak drain voltage. The peak reverse recovery time and the parasitic capacitances of the SRFET also affect the peak drain voltage of the primary. You might want to look into all these things and compare it with DER494 to get a performance comparable with it.

PI-AP

Hi PI-AP,

Thanks for the suggestion of the application example. I think your DEE533 really meant DER533. It looks a good example for PD2.0, but I really need latest design that is fully compliant with PD3.0, which should supports up to DC 20v 100w output. That is not to say I must build a PD3.0 adapter with INN2215 to output 100w, but with the same architecture, we can easily scale it up to 100w with minimum circuit modifications. If 100Watt is really that hard to achieve, can we do at least 70w, which would allow the adapter to be applicable to most laptops?

Back to your questions. When the current go over 0.95A at 12v, the power supply appear to power off it self, which gives no output and could be recovered only by recycling the AC input. For the scope captures, I think that I must have used AC-coupling, and the probe was set to 10x, ie. actual scale was 100v/div. Are these captures invalid?

The PI expert is really a convenient design tool. I have already started to use it. for my circuit, I have put totally 33uf for the input filter cap, ie. C1+C2 in my sch. It sounds strange to me that this circuit can output more than 4A at 5V but could not go over 1A at 12V. Is there any hint in the PI expert that can help to find the cause of the problem? I posted previously scope captures of 12v 0.95 A output, does that good or not?

My snubber uses a RS1J, the datasheet is attached. Is it suitable to be used here? What values are necessary to be changed for snubber using this diode? Lastly, is the high drain spike a cause for the problem of low output current at 12V? 

Lastly, I seem to have no tool to measure the leakage inductance, especially at 100khz, like what suggested in the DER494. I only have a 5MHz function gen and a 100MHz DSO. Any clue to use these to measure inductance?

Thank you! Have a nice day.

Jack

Hi Jazzclassics,

PD3.0 compliant 100W design is in progress. DER533 is the only USBPD design published yet. We have 27W and 45W design also in progress.

The power supply seems to latch off which could happen if the current being pushed into the BPP pin is close to 9-10mA. Can you please measure this and the bias winding voltage just before the latch off?

The scope captures do not show a real time scenario magnitudes for the drain voltage if we go by calculations, so I was asking about its validity by checking the coupling/attenuation factor.

The PI Expert transformer design should show warnings if it would get power limited for certain designs for eg: the 12 V design in your case.

To reduce the peak drain voltage, you can try and increase the capacitor C4 or decrease R1||R2 or derease R3 and see the effect on the peak drain voltage magnitude. You are using a fast diode now, you could replace it with a standard diode to reduce the peak drain voltage further.

PI-AP

Hi PI-AP,

I was looking for this reply for all these days. Thanks a lot.

I measured with a DMM the voltage across the R5(22ohm). I have got the result as the follows. The IR5 is calculated by dividing the VR5 by the resistance 22ohm. Is this a valid way of calculating the current through BPP? I have forgot to measure the bias winding voltage, will do it next time I have got the time. But could you tell from this table that whether the current through BPP pin is too large to cause a latch off? What else could be the cause of the latch off?

I also recaptured some drain waveforms by using DC-coupling, pictures attached. I have decreased the R3 by parallel it with another 270K, and increased C4 by parallel it with another 1nF. But none of these bring significant change to the high drain spike. Please have a look a the capture and let me know your thinkings.

Please try help to reply asap.

Thanks a lot

Hi PI-AP,

Good to hear that you have PD3.0 compliant design in progress. Before its release, could you recommend some references that is useful for building an PD3.0 adapter? Thanks a lot.

Hi Jazzclassics,

I apologize for the delayed response. Please find attached the 'Primary Bypass pin shutdown threshold current' value from the Innoswitch CP datasheet. The current value you are reporting is within the max and min range and hence is leading to latch off. You need to measure the bias winding voltage on your board for the min and max output voltage conditions and adjust your zener ZD1 votlage rating or resistor R4 and R5 values such that no more that ~1.1 mA of current flows to the BPP pin under normal operating conditions.

The next way to reduce your drain spike is by reducing the leakage reactance on your transformer by changing the way you are winding it as the drain spike results due to the leakage reactance. I am not able to know the actual peak voltage on your drain waveform as you are on 10V per division and by that it comes to only 40V- 50V which is impossible. So please use the right probe/oscilloscope settings to capture the right peak drain voltage magnitude. Also it will help if you callibrate your probe before measuring these waveforms.

PI-AP

Hi PI-AP,

It is so good to get your help. The BPP pin current is close to latch off range before the shutdown. 

For the drain waveforms, my DSO voltage measurement upper limit is 400v, but it has to display it under 40v range by using a 10x probe, which appearantly divids the input voltage to 1/10, this is what told by the manual. Therefore the 10v grid displayed actually stands for 100v. Does this make any sense?

I suspect this is a problem with the transformer, ie. the supplier does not make it right, high leakage inductance, high bias voltage, and inferior core etc. but have not figured out a way to proove it.I keep an extra few unusef transformers samples of the same batch, is there a way to measure the inductance of those by using a DSO and a function generator?

Now that I have to be busy with a 80v 6A half bridge converter project, it has to fit in a 20x10x5cm housing. I tried the PI expert but it looks that does not  provide a solution for converter other than flyback topology. Do you have any reference desigh for such? The initial design has choosen SG3525, the datasheet says it could switch up to 400kHz. I need at least 120kHz. I am not sure if this is a right choice.

Today is Friday, have a nice day and good weekend :)

Hi Jazzclassics,

We do not know any way to measure the leakage inductance using DSO and function generator. I am unable to comment if your waveforms are right or not as I do not know your input voltage and assumed VOR for the transformer design.

HiperLCS is our product which is used for half bridge converters. PI expert does not have it but PI Xls has it. You could use DER385 as reference design. Our website has other reference designs as well for this product. This product doesn't need an external driver as it has everything integrated in it.

PI-AP

Dear, PI_AP. and PI_ODO,

Firstly, would like to Express my sincere appreciation for your help in the previous posts.

I have replaced the transformer.
with a transformer from another supplier, that has resolved most of the problems. That teachs me a lesson that don't start implementation without knowing what your transformer should like.

Secondly, I have queried you, at half year ago, about the higher power AC DC USB PD compliant adapter design up to 100 watt. And I have got the answer that Such a design were undergoing and you were going to release that. And today, I have got from the website that you have, just a few days ago, released the inno-switch3 series, which will facilitate USB PD adapter design up to 55w.Congratulations !

I notice that the new innowhich3 has about the same specified primary peak current as the old innoswitch CP chip. Why Can't we use old innoswitch CP To make adapters to deliver 30+ or even 50+ watt power? I'm still debugging a design with the innoswitch, but you already turning on your heel and heading for innoswitch3.

In your design example DER567, You have used the WT6630 in an adapter up to 9V2A. Is it possible to modify the design and make it 20V1A? What changes needed to be made to this design to achieve a higher voltage output ?

By the way, there is an error in your Web page.

Dear jazzclassics,

Thank you for the kind comments! We are very excited about our InnoSwitch-3 launch! The InnoSwitch-3 is able to support the higher power levels, as the larger-sized devices in the family have higher current limits (please see corresponding Ilim numbers in the electrical tables). It also helps that the InnoSwitch-3 family operates with higher efficiency. By the way, thank you for flagging the power rating error on our web page.

Regarding DER567, are your modifications for 20V1A for a USB-PD power supply that will also need to operate at lower voltages such as 5V and 9V?

-PI-ODO

Dear PI-ODO,

Thank a lot for your reply. It will need to out put 5v 3A, 9v 2.2A, and 20v 1A. Is this doable with a inno-switch CP 2215k?

Dear PI-ODO,

I have captured some transformer primary waveforms at full load. From those with forms and I can see that the mosfet does not switch on all the times When it outputs 9.5 volt to a 2.35A load . There's a gap about 50us between the pulses. And even when overloaded at 9.6v to a 2.7A load, there is still a gap.
Please comment on this with waveforms.
Is it possible to optimize this design to output higher power?

In the INN2215k data sheet, it says The drain pin peak current is 1680ma, but there is a scoped 3150, what does this 3150 stand for?

I have calculated that I only need1.64A peak primary current for a 60W output, where did i go wrong?
VACMIN=110v
VACMAX=265v
FL=50hz
fS=100000hz
VO=20v
PO=60W
n=0.86
DMAX=0.46
KP=0.4
TRANSFORMER
BM=0.18
AE=0.86cm^2

Ip=2*PO/(n*(KP+1)*VACMIN*1.2*DMAX)
=2*60/(0.86*(0.4+1)*110*1.2*0.46)
=1.64Amp

I picked up a EE28 core for the prototyper, hoping it be able to scale up to 60W. Is it realistic to make it 60w with a innoswitch3?

Dear jazzclassics,

To support 20V in addition to 5V and 9V in a single output, the design will require updates to the transformer, SR FET selection, and primary clamp. For specific details, we recommend that you use the PI Expert / PI Xls design tool to re-calculate the adjustments needed. In particular, the InnoSwitch-CP and InnoSwitch3-CP tools allow a user to specify a set of output voltages, enabling you, for example, to add the case for 20V/1A. Based on our experience, 20W should be doable with the InnoSwitch-CP. Beyond that, we would recommend moving to the InnoSwitch3-CP, allowing for power levels up to 55W. 60W may be doable with the InnoSwitch3-CP, but it is on the cusp, depending on thermals, etc.

Regarding the drain pin peak current, the number in parentheses describes the momentary, higher-level current allowed during transients, when the drain voltage is simultaneously below 400V. The higher rating is not intended for sustained cases.

Finally, in regard to the waveforms, they appear to be consistent with what is expected near maximum loading (see attached picture).

-PI-ODO

Dear PI-ODO, Thanks a lot for your reply. Where can I get samples of innoswith3? I have followed your design example DER567, but had problem to get the PD part to work with any of my test loads. It appears that the WT6630P is not working. There is no output at the VBUS as a result that the Q1 and Q2 are not switched on. The DC voltage measured at the postive pin of the C3(470uf 16v) is 9.2v at no load. And I short the WT6630P gate pin(pin3) to RTN(GND) to find the 9.2v appear at the VBUS, which proves the Mosfets are good, but the PD interface remains irresponsive to any PD requests.I have replaced the WT6630P and the associated circuits with another PD interface board and find that the whole circuit is working well, which can out put 5V, 9V, 12V at 3.2A, 2.6A, 1.8A respectively. 

This puzzles me a lot. Why the PD interface is not working? And if it not worked the voltage at C3 should be 5v instead of 9.2 v, correct? It being 9.2v means that the WT6630P sinks some current through GPIO1/R16, correct? 

Please help!

I need more details of how to debug the PD interface part. My email address is jazzclassics@163.com.

thank you.

Dear jazzclassics,

I will ask someone on our team to reach out to you at your email address. For InnoSwitch-3 samples, I recommend you take a look at the following link, where you can specify and order what you want: https://ac-dc.power.com/products/innoswitch-family/build-your-own-innoswitch/

-PI-ODO

Dear PI-ODO,

Thank you for bringing me your team mate.

I am in Shenzhen China. Maybe you can let me contact your local tech support?

Thanks again for your help.

Dear jazzclassics,

Here is the contact information for our Shenzhen office:

China (Shenzhen)
17/F, Hivac Building, No. 2, Keji Nan
8th Road, Nanshan District,
Shenzhen, China, 518057
Phone: +86-755-8672-8689
e-mail: chinasales@power.com

They can help to refer you to their local tech support.

-PI-ODO

Hi..as per my knowledge the INN2215K SR MOSFET driver typically uses a supply rail of 4.4 V, a MOSFET with too high a threshold voltage, VGS(th), is not suitable. It is recommended that the SR MOSFET have a threshold voltage well below the drive level. See Figure 1 for preferred MOSFET characteristic.

Hi NoellEagan,

I am using AO4290. Is this fet a good choice?

Thank you very much for your advice.

Dear PI-ODO,

Could you send me a detailed datasheet and setup process for WT6630 used in your DER567? I have contacted the PI support engineer, but he did not have that file.

Thank you very much.

Dear jazzclassics,

Per our partner Weltrend, please refer to the “Support” section on Weltrend’s website to download related app notes or user manuals.

You can also pass questions to Weltrend by going to the “Contact Us” link of their website:

http://www.weltrend.com/en-global/about/index/18

Finally, you can also reach out to Weltrend regional sales contacts, and they will answer your questions as soon as they can.

-PI-ODO