Questions about RDR203

1..Why not use a mains voltage doubler link?
>You can but most customers prefer a true universal input

2...Why is the common mode choke not rated for the peak RMS current from the mains?
>Because they are thermally limited, not peak current limited.

3...How long can this pcb survive peak load at min vin?
>Until it shuts down from OTP which can take several minutes, depending on your ambient temperature.

4...Why did you not use PWM control?....surely on/off control is too noisy at 100W for a flyback?
>For the same reason we have sold 100's of millions of TNYSwitch - it enables a low cost system solutions. Noise is a non-issue.

5...Why are the output feedback divider resistors made of metal film, when most of the other resistors use carbon film?
>1% tolerance. You are welcome to use 5% if your output voltage tolerance spec is loose enough.

6...Why is the NTC not rated for the mains rms input current that occurs at peak load/min vin?
>Because they take several seconds to a minute to overheat. Peak load requirements don't drive its thermal spec vis a vis the spec.

7....If the flyback transformers are NOT varnished....will there be audible noise?
>Possibly.

8....If a 100W SMPS can be done with simple on/off control like rdr203, then why doesnt on/off control get wider use at this power level.
>Audio is unique with its very low average to peak power ratio.

9.....All i understand is that EN pin sourcing >240uA means the FET stops switching....what is the mechanism of the current limit adjustment?...i could not ascertain this from the datasheet?
>If you are asking about the Drain current cycle-by-cycle current limit, it is internal to the controller.

10....How can you be sure that the audio amplifier is not a class d guitar amplifier?...due to being ovedriven, these can give peak load all the time.
>This depends on your specification. RDK-203 was designed for the typical home audio amplifier application. You are welcome to test the DRK-203 to a guitar amplifier specification.

11.....the overload protection is disabled in rdr203...so what happens now when there is a continuous overload?
>If the overload is enough to heat the device to the OTP threshold, it will shut down eventually. It can take up to 10s of minutes with severe overload. The time will depend on your ambient temperature.

12.....pks607y runs at 277khz at peak load.....thats above the 150khz threshold frequency of conducted emissions...so why use it?
>The EMI filter is oversized and shows large margin - if EMI is tested in the system with an actual amplifier playing actual music the EMI will be low.

13....PKS607Y rds(on) is 2r5, surely this cannot support the 100W loading of rdr203 for long?
>The report shows actual thermal and other test results.

14....Where did you calculate your mains harmonic current magnitudes?.....are you aware that even though PFC is not required for audio, you still have to conform to stringent levels of mains harmonic currents?.....why are mains harmonic currents at peak load not calculated.
>PF is not measured at peak load but during actual system loading.


15....why did you not just use a cheap boundary conduction mode boost PFC stage followed by a single on/off control flyback with split rail outputs?...your primary current would have been far lower, and transient response even better.
>There is always an engineering cost tradeoff in meeting all the specs and this design meets the specs it was designed to meet at the time it was conceptualized. It makes no sense to get a customer to pay for a PFC if it is not needed, "just to lower input current", when the input power when playing actual music will likely be <15W even at maximum volume.

16….what did the radiated emissions scans look like at pk load /min vin?
>Specifications don't expect measurements at peak load. They are done during actual loading of actual usage. Radiated EMI is very dependent on the way the system case is designed and how the boards are grounded to the case. And radiated EMI issues are usually fixable without adding significant cost, and usually with no cost solutions (e.g. positioning of grounding points, etc)

17….what did the common mode conducted emissions scans look like at peak load/min vin?
>Specifications don't expect measurements at peak load. They are done during actual loading of actual usage.

18…What are the mains harmonic current levels at pk load/min vin.?
>Specifications don't expect measurements at peak load. They are done during actual loading of actual usage. When this board was conceptualized, there were no requirements for line current harmonics. The average input power is low.

19….will you get nuisance tripping of the mains fuse?…even with the ntc.
>No, this board has been tested using an actual amplifier load playing actual music with lots of dynamic compression, e.g. FM radio.

20…..wont the input electrolytics be damaged by the inrush at 265VAC…even with the ntc?
>Electrolytics can take an enormous amount of peak current. It doesn't faze them.

21…..why did you bother to disable the overload protection with the 555 timer…..in overload, the bias rail will drop out, and disable the controller anyway.
>In overload, the bias rail goes *up*, not down, due to the effect of transformer leakage inductance.

22….Why didn’t you just use a controller with no overload protection…eg UCC28C42…..instead of using PKS607Y and then having to disable the overload protection.
>Because the PKSwitch with its integrated MOSFET is a lower cost solution.

23……UCC28C42 reduces the current limit when at light load……so why is the pks607y better than ucc28c42?……what is its selling point?
>Because the PKSwitch with its integrated MOSFET is a lower cost solution

24…how do you ensure that you don’t get resonation in the dual stage emi filter?
>From actual EMI testing.

25…..why didn’t you use surface mount components anywhere…it would have made noise-free pcb layout more possible.
>Cost. Many customers prefer a no-SMD solution. Noise is not an issue with careful layout.

26…..Is it not possible for a PWM controlled solution to give sufficient transient response for this application?
>That is up to you and your spec.

Re-visiting your answer to Qu 15, -you say the power is generally 15W.....so then why have you used TWO flybacks?
15W is well within the capability of a SINGLE flyback.

And regarding "peak power" in audio applications......i've never known an audio engineer who can actually say what the longest time duration of peak power is.........you ask them "is it milliseconds"......"is it minutes?", "is it hours?" and they just can't answer...........so therefore you have no choice but to do an smps that can handle the peak power rating CONTINUOUSLY....surely you agree with this?

No I do not agree. Such a spec will result in costly over-design. Ask them what the amplifier full power testing procedure is, and test the PSU with that amp and that test. If that test is unrealistic vis a vis music (e.g. sine wave tone just below clipping, for 10 minutes), ask them if it is acceptable that the PSU output votlage droop during that (unrealistic) test. One may be able to design an SMPS and amplifier that passes those tests (e.g. no thermal shutdown) but may produce paper power ratings that are below what is realistic in terms of reproducing music. That is, the actual music peak power with no distortion, may be much higher than what the numbers derived from old test standards suggest.

Bear in mind that some of the IEC style audio power amp rating testing was designed with the old 60 hz PSU's and class AB amplifiers, 30 years ago. Many of those tests are obsolete in light of SMPS and class D amps.

15W may size the thermals, but the PSU must have a peak power capability equal to the max power averaged over a 1~5 ms period, else the output will droop.

ok but i dont know where you are getting the 1-5ms period from?, -how do you know that the peak power is not continuously experienced for say several minutes?
And also, a "strum" of a guitar gives peak power, and this strum can last for 250ms and be repeated continuously.....so then you will be much above your 15w average figure....surley you agree, that there is no "normal" music for guitar?

..the strum is the peak power, and can be continuos strumming for long time....and then RDR203 will shut down with OTP and the audience at the gig will want there money back....surely you agree?

ok but i dont know where you are getting the 1-5ms period from?, -how do you know that the peak power is not continuously experienced for say several minutes?
And also, a "strum" of a guitar gives peak power, and this strum can last for 250ms and be repeated continuously.....so then you will be much above your 15w average figure....surley you agree, that there is no "normal" music for guitar?

..the strum is the peak power, and can be continuos strumming for long time....and then RDR203 will shut down with OTP and the audience at the gig will want there money back....surely you agree?

1-5 ms is the time period over which an SMPS output may droop significantly when overloaded. It is the "holdup time" if you will, of the output capacitor. IOW with a peak load of greater than the PSU peak load, the shortfall can be provided by a large output capacitor. A 4,700 uF capacitor, for example, has a time constant, with an 8 ohm speaker load, of 30 ms.


Remember that for a sine wave (or music) signal, the load on the PSU is the full wave rectified version of this signal. For a 100 Hz tone, the half sinusoids are 5 ms wide. The peak of this sinusoid can be partially supplied by the output capacitor.


Enlarging the output capacitor to provide the "ride-through" for very short peak load transients can be less costly than increasing the peak power of the converter.


RDK-203 was designed for a home audio application. For a guitar amplifier application, examine the spec closely and compare it against the RDK-203 specification. Remember that a lead guitar's lowest tone is 83 Hz and a bass guitar is an octave lower. From the specification, you can determine RDK-203's suitability. Better yet, test the actual unit against the guitar specification. Watch for the output drooping, and log the PKSwitch temperature.

I am sorry but using a sine signal as an example as not accurate…..there can be many of these simusoids at different phases…..and the effect is peak load all the time.

Treez, look at an actual music waveform.
Then please re-read what I wrote.

Now think - how would the music waveform load the PSU?

If you want to carry the discussion further along these lines, pls. go ahead and test an RDR-203 with a guitar.