PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2021-12-07, 21:15:41
News: Check out the Benches; a place for people to moderate their own thread and document their builds and data.
If you would like your own Bench, please PM an Admin.
Most Benches are visible only to members.

Pages: 1 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 [25] 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 ... 103
Author Topic: Dally, Shark & Ruslan workbench  (Read 76272 times)

Hero Member
*****

Posts: 749
Vasik,
i did just that, "simultaneous winding with two wires of equal length."
I will have to see if i can make them equal induction wise.
Itsu

Itsu,

that is strange, why would they have different inductance ?
BTW have you checked that delays (RV1,5 C19,20) equal for both push pull legs ?

Vasik


---------------------------
The only way of discovering the limits of the possible is to venture a little way past them into the impossible.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

Hi Vasik,

There always will be some difference between the both primaries i guess, only by how much is still acceptable.
I will examen the primaries to see if i made some mistake somewhere or have one side against/over the gap (the both primaries just fit the one half of the yoke.

Below sreenshots are the TL494 outputs (yellow and blue) together with their delayed (gate) signals (purple and green) :


Itsu


   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Itsu,

Still continues to make no sense to me.

Before tearing into the coil, I would replace the yoke windings with a pair of resistors again and confirm all is as it should be. 

The 48vdc at the drain makes no sense to me, nor does the purple drain being pulled low when both the blue and yellow gate are low. 

What happened to your current probe?

PW
   

PW, 

i put some 470K resistors at the primaries position and measured the gate and drain signals (green and purple gates, blue and yellow drains).

Using still the single 24V PS, and this shows as expected.


I somehow managed to short out the current probe insides to a life wire which i was measuring, and it blew up the Hall Effect device so no DC currents are being recorded, just AC.


Itsu

   

Hero Member
*****

Posts: 749
Below sreenshots are the TL494 outputs (yellow and blue) together with their delayed (gate) signals (purple and green)

Itsu,

Delays look different even with this small scale.
This could potentially cause that core magnetized in one direction more than into other
and create dis-balance you observing.
Please re-check the delays.

Vasik


---------------------------
The only way of discovering the limits of the possible is to venture a little way past them into the impossible.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

Vasik,

i agree, they look different, but my thoughts were that the duty cycle (all around 33%) would be the major factor for heating up.

But i have adjusted the delays to be almost the same (13.8°), see screenshot yellow and blue the TL494 signal, green and purple the delayed (gate) signals.

Still the one MOSFET gets hotter then the other (26°C v 41°C in 5 minutes, only primaries connected).

So i will examen the primaries for obvious unbalans.


Thanks,   Itsu 
   

Hero Member
*****

Posts: 749
Itsu,

I agree, it is good idea to check primary windings also.
If you have current probe broken, you can make one yourself from a small ferrite core and resistor.
You wind 20 turns and adjust sensitivity with resistor. You can calibrate it with signal source and some known resistor.

Vasik


---------------------------
The only way of discovering the limits of the possible is to venture a little way past them into the impossible.
   

Group: Professor
Hero Member
*****

Posts: 2602
I somehow managed to short out the current probe insides to a life wire which i was measuring, and it blew up the Hall Effect device so no DC currents are being recorded, just AC.
Were you able to identify the part number of the Hall sensor inside?
   

Group: Professor
Hero Member
*****

Posts: 2602
If you have current probe broken, you can make one yourself from a small ferrite core and resistor.
You wind 20 turns and adjust sensitivity with resistor. You can calibrate it with signal source and some known resistor.
His half-broken current probe is superior to the core+resistor probe.
Both don't sense DC.
   

Group: Professor
Hero Member
*****

Posts: 2602
Is the shorter period of oscillations that occur during both gate off times the beginning of clamp diode turn off?
Yes and/or the body diode of the opposite mosfet.

The two screenshots below illustrate how the drain waveforms appear without* the body diodes in that other split-primary push-pull experiment:

  • The 1st scopeshot shows how the drain voltage waveforms (yellow & purple) appear with the body diodes present.
  • The 2nd scopeshot shows how the drain voltage waveforms (yellow & purple) appear with the body diodes "absent"*.

( in both cases the drain waveforms are clamped to ±70V by bilateral TVS diodes )
( the flux in the core never becomes discontinuous because of the high duty cycle )
( the gate waveforms are blue and green )


P.S.
Even on the 1st scopeshot, you can see that the drain voltage is 1 pixel lower when these body diodes conduct, if you look very closely.  That's the -1.2V forward voltage of these diodes at low vertical magnification.  At less V/div and very high acquisition averaging, this could be seen even more clearly.




* achieved by connecting two MOSFETs in series, so their body diodes are mutually opposing.
« Last Edit: 2021-05-24, 13:18:10 by verpies »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Were you able to identify the part number of the Hall sensor inside?

Yes,  its this part:

I have one on order.


Itsu
   

Group: Professor
Hero Member
*****

Posts: 2602
Yes,  its this part:
I have one on order.
Looks expensive  :(
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

I rearranged the yoke wiring and was able to get the both primaries fairly equal, as both measure now around 42uH.
I also shortend the yoke leads to the pcb which cleaned up the signals.

With only the 5 turn secondary connected, i have an input of 24V @ 700mA.

Temperature on the MOSFETs are stable around 58 and 55°C.

When i connect the other secondary (18 turns going to grenade with parallel 300nF caps and the output circuit), the input current increases to about 8A @ 24V!

Swapping this secondary coil leads makes no difference.


Screenshot shows the MOSFET 1 current (green), drain signal (yellow), gate signal (purple) and the MOSFET 2 drain signal (blue).

Video here: https://youtu.be/DSy4wE09H-U

Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

Here i have changed the current probe position to the 5 turn secondary circuit and tuned for max current (green) being resonance at 111KHz which is the 5th harmonic of the TL494 base frequency at 22.46KHz.

As i aim for a base frequency of 24KHz i have to lower the Inductor series cap (now 27nF) a bit to get to a resonance frequency of 120Khz in the inductor.

Itsu
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
Itsu,

Judging by your drain scope traces, both mosfets are in avalanche for a short duration which can add considerable dissipation.

Regards,
Pm
   

Group: Professor
Hero Member
*****

Posts: 2602
Judging by your drain scope traces, both mosfets are in avalanche for a short duration which can add considerable dissipation.
Do you mean, right after the gate signal goes low?
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

I also connected up the 2nd secondary (18 turns across the primaries) which goes to the grenade and 300nF parallel caps and the output circuit with a 40W / 220V bulb.

The input went up to almost 8A @ 24V and we have ~290V DC on the bulb.

Temperature of the MOSFETs increase quickly, but might need some tuning done (to 24KHz).

Screenshot shows the current in the 2nd secondary parallel circuit (green), the MOSFET 1 gate signal (purple) and drain signal (yellow), and the MOSFET 2 drain signal (blue).

Video here:  https://youtu.be/WIraRkX2Bc8

Itsu
   

Sr. Member
****

Posts: 340
I rearranged the yoke wiring and was able to get the both primaries fairly equal, as both measure now around 42uH.
I also shortend the yoke leads to the pcb which cleaned up the signals.

With only the 5 turn secondary connected, i have an input of 24V @ 700mA.

Temperature on the MOSFETs are stable around 58 and 55°C.

When i connect the other secondary (18 turns going to grenade with parallel 300nF caps and the output circuit), the input current increases to about 8A @ 24V!

Swapping this secondary coil leads makes no difference.


Screenshot shows the MOSFET 1 current (green), drain signal (yellow), gate signal (purple) and the MOSFET 2 drain signal (blue).

Video here: https://youtu.be/DSy4wE09H-U

Itsu


    You did not mention the temps of the Fets at almost 200w input, without any load on. That's what I need to know. Sorry to ask...
    One of your fet's output is still a little higher than the other. What are the temps of your fets at resonance of the push pull with the 18t secondary coil to the grenade connected up and drawing 8 amps? But, please don't blow your fets up on my account. I know that can happen, especially without any fans on. Keep your finger on those fets....We're going in...

   NickZ
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
Right after the gate signal goes low?

Yes, from the collapsing field of each primary winding.  Need to lower the supply voltage or use mosfet with higher BVdss rating.

Pm
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
I also connected up the 2nd secondary (18 turns across the primaries) which goes to the grenade and 300nF parallel caps and the output circuit with a 40W / 220V bulb.

The input went up to almost 8A @ 24V and we have ~290V DC on the bulb.

Temperature of the MOSFETs increase quickly, but might need some tuning done (to 24KHz).

Screenshot shows the current in the 2nd secondary parallel circuit (green), the MOSFET 1 gate signal (purple) and drain signal (yellow), and the MOSFET 2 drain signal (blue).

Video here:  https://youtu.be/WIraRkX2Bc8

Itsu

Itsu, these traces show a considerable avalanche in each mosfet.  This is going to create considerable dissipation and heat.

Edit: If you take a pix of the drain voltage and drain current, the energy dissipated per each mosfet can be calculated and then the overall power can be calculated at the operating frequency.

Pm
   

Group: Professor
Hero Member
*****

Posts: 2602
Need to lower the supply voltage or use mosfet with higher BVdss rating.
...or increase the load on the secondaries so the magnetic flux energy, which is stored in the core, has other places to go rather than breaking down the D-S junction of the MOSFET, after the gate goes low.
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
...or increase the load on the secondaries so the magnetic flux energy, which is stored in the core, has other places to go rather than breaking down the D-S junction of the MOSFET, after the gate goes low.

Yes, this will work if the coupling is tight between the secondary and primaries.  However, my guess is that the coupling with the yolk core is not that great so even with the secondary shorted, there would probably still be evidence of avalanche.

I'm not that familiar with the overall operation so I'll ask, is the output of the squarewave of the yolk transformer converted to sine somewhere before being used?

Pm
   

Group: Professor
Hero Member
*****

Posts: 2602
Yes, this will work if the coupling is tight between the secondary and primaries.  However, my guess is that the coupling with the yolk core is not that great so even with the secondary shorted,
You can estimate the coupling coefficient by the fact that the inductance of his primary is 42μH when the secondaries are open and 4μH when the secondaries are shorted.

there would probably still be evidence of avalanche.
It would be nice to have a reliable indicator of the dangerous D-S voltages. MOSFETs are not cheap and they can often fail only partially, leading to malfunctions which are difficult to debug.

An overvoltage detector/protector can be made by putting a TVS diode across the drain & source ( a TVS diode with a breakdown voltage rating close to MOSFET's BVdss ) and surrounding it with a small toroidal core and an LED like on the diagram below*.



This overvoltage detector/protector does not affect the operation of the circuit when the current never flows through the TVS and the LED never lights up. 
It is just a warning light ...and lighting it up should be avoided during normal operation.


* a bridge rectifier made out of 4 small (but fast) diodes, followed by a cap in parallel (e.g 10μF) and placed between the winding and the LED, would make this indicator much more sensitive and easier to see.
   

Hero Member
*****

Posts: 749
I also connected up the 2nd secondary (18 turns across the primaries) which goes to the grenade and 300nF parallel caps and the output circuit with a 40W / 220V bulb.

The input went up to almost 8A @ 24V and we have ~290V DC on the bulb.

Temperature of the MOSFETs increase quickly, but might need some tuning done (to 24KHz).

Screenshot shows the current in the 2nd secondary parallel circuit (green), the MOSFET 1 gate signal (purple) and drain signal (yellow), and the MOSFET 2 drain signal (blue).

Video here:  https://youtu.be/WIraRkX2Bc8

Itsu

Itsu,

What is voltage on inductor ?
My understanding is that it should be around 200v
and same should be output of gradient coil.
If voltage too high you might need remove some turns from yoke secondary windings.

I would suggest tuning inductor and gradient coil separately.
When you come close to 24.4Khz you should see HF harmonics, this will help you with precise tuning (look for highest HF signal).

Vasik


---------------------------
The only way of discovering the limits of the possible is to venture a little way past them into the impossible.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

    You did not mention the temps of the Fets at almost 200w input, without any load on. That's what I need to know. Sorry to ask...
    One of your fet's output is still a little higher than the other. What are the temps of your fets at resonance of the push pull with the 18t secondary coil to the grenade connected up and drawing 8 amps? But, please don't blow your fets up on my account. I know that can happen, especially without any fans on. Keep your finger on those fets....We're going in...

   NickZ

Nick,

those MOSFETs get hot quickly then, like within half a minute they are 70°C and rising, but look at the drain signals, they avalange? now twice/cycle.
Need to tune (frequency / resonance / duty cycle / number of turns) to see if i can improve.

A perfect match between the MOSFETs will be very hard to get i think.

Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Itsu, these traces show a considerable avalanche in each mosfet.  This is going to create considerable dissipation and heat.

Edit: If you take a pix of the drain voltage and drain current, the energy dissipated per each mosfet can be calculated and then the overall power can be calculated at the operating frequency.

Pm

PM, 

i understand this "avalancing" is kind of needed in this circuit, there is a PDF/video on it which i need to follow to get it right (frequency/Duty cycle).

But it should be below the max parameters of the used MOSFETs.


Itsu
   
Pages: 1 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 [25] 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 ... 103
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2021-12-07, 21:15:41