Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

TC4422 question re-phrased

Status
Not open for further replies.

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
The TC4422 chip is in place of the BD140/BD139 totem pole. Unfortunately I do not have a multisim model for this chip so I can add it to the circuit below.

There is also 4 series 1N5819 schottky diodes across the mosfet to protect it.

Unfortunately I appear to fried my current TC4422 as it is powered from the 12V rail, rather than the 9V rail, and is therefore not potected by the 1N4745 schottky and 470uF capacitor. I over looked this when I was soldering it up.

Question is what would be the most appropriate protection for the TC4422.

Would a 1n4007 or similar on the two Vcc pins of the TC442 be sufficient to prevent flyback transient voltage spikes going through it?

Or would a capacitor/resistor snubber with a reverse biased 1n4007 be better to provide protection for the TC4422?

Or both?

By the way the TC4422 works briliantly so thanks to those who pointed me in the direction of these.

With my crappy old battery the power mosfet did not get noticeably hot while my jacobs ladder was running, where as with the transistor totem gate pole driver it did get quite hot. The mosfet did start to get a little hot when I connected my circuit to a better car battery so it may well have melted down if tried to run it with the totem pole driver and the better battery. I have a fairly large heat sink on the mosfet anyway.
 
Last edited:

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
In this circuit I have a TC442 between the 555 and the mosfet but I can't add it to this circuit schematic as I cannot find a TC4422 model for multisim so you will just have to imagine it is there.

What would cause this chip to fail when the above circuit is operating off a good car battery with perhaps 3A or so going through the tv flyback transformer, but it doesn't fail if the same circuit is operating off an old car battery that probably would be lucky to output 1A.

I confirmed that it was the TC4422 that failed because I have arranged the circuit with connectors that allow me to divert the output of the 555 to an LED and also divert the output of the TC4422 to the same LED.
Replacing the 555 timing capacitor with 1000uF I can then see the LED flash.

So after the TC4422 failed and my jacobs ladder stopped working the 555 was still vibrating OK but the vibration was not passing through the TC4422. I replaced the TC4422 and the vibration successfully passed through the TC4422.

So what is likely to have killed the TC4422?
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
Well actually I have reasoned that, since the jaycar jacobs ladder circuit used a 1N4007 on the supply rail for the 555 and which I have emulated in my circuit here, then it should be enough to put 2 x 1n4007s between the Vcc pins of the TC4422 and the 12V supply rail.

Is that sound reasoning do you folks think?
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
I see two points:
- simulation. A TC4422 Spice model is available from Microchip and can be imported in any Spice based simulator, also Multisim
- hardware. As far as I understand, the real problem is failure of your hardware design. I won't expect any insights in this regard from a simulation, because you don't see the effects that possibly damaged the device.

As a side remark, I don't understand the power circuit as shown in post #3. Unfortunately it can't be viewn in detail. Do you have Z-diodes in parallel to the MOSFET?
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
I see two points:
- simulation. A TC4422 Spice model is available from Microchip and can be imported in any Spice based simulator, also Multisim
- hardware. As far as I understand, the real problem is failure of your hardware design. I won't expect any insights in this regard from a simulation, because you don't see the effects that possibly damaged the device.

As a side remark, I don't understand the power circuit as shown in post #3. Unfortunately it can't be viewn in detail. Do you have Z-diodes in parallel to the MOSFET?

I have taken the jaycar jacobs ladder circuit, which I purchased and which uses a darlington rather than a mosfet, and incorporated the protection feature of that circuit into this mosfet version.

It uses 4 series schottky diodes in parallel with the darlington, or in this case the mosfet.

jacobs_ladder_schematic_sm_sc0407.gif


As you can see the darlington is driven by a BC327 and the BC327 is protected from transients spikes via the schottky and capacitor within the 555 branch of the supply rail.

In my case the mosfet is driven by the TC4422 and that is suplied from the 12V rail and is therefore outside the protection for the 555 etc.

So if a 1n4007 protects the 555 etc from reverse transient spikes then 1n4007 should also protect the tc4422 from reverse transient spikes.
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
I don't want to guess about the failure mechanism. But 1N5819 schottky diodes aren't a suitable replacement for 75V Z-diodes.
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
I don't want to guess about the failure mechanism. But 1N5819 schottky diodes aren't a suitable replacement for 75V Z-diodes.

Sorry ignore those 1N5819, I dunno what I was thinking when I added those to the scematic. In the real circuit they are zenner diodes.
 

mikhelee

Newbie level 1
Joined
Sep 26, 2012
Messages
1
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,284
Yes, i resolder it again in couple of days, see if problem occurs and if yes i will try to examine signals. I'll be probably back in 4-5 days. Thanks for suggestions.
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
In this circuit the 2 x 1N4007s do not protect the TC4422 from transient spikes - it works for a short while before the arc stops along with any oscillations fro the TC4433.

In hind sight this makes sense since the current direction will the the same through the TC4422 no matter which way it is flowing along the 12 supply rail. Hence the 1N4007s cannot protect the TC4422 from the spikes.

I have since added a snubber according to thei website: https://www.daycounter.com/Calculators/Snubbers/Snubber-Design-Calculator.phtml

I found an inductance value for a flyback transformer of 1uH. Is that likely to be at least ball park for a flyback from a tv?

I am unsure however as to whether the snubber should be connected across the primary coil of the flyback or whether it should be connected across the 12V and GND rails, or whether it will make any difference.

Any help would be appreciated.
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
In this circuit the 2 x 1N4007s do not protect the TC4422 from transient spikes
In "this circuit", there's only one 1N4007 and no TC4422. Looks like a riddle "guess the original circuit", maybe posted in the wrong forum?
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
In "this circuit", there's only one 1N4007 and no TC4422. Looks like a riddle "guess the original circuit", maybe posted in the wrong forum?

In my browser the circuit shows with the dummy TC4422 that I added to it. Would it be a case of you needing to delete your browser data and refresh the page?

The TC4422 doesn't work when you try to run the simulation as there is some incompatibility with the model data and multisim 11. But it is good enough for the purposes of showing the schematic.

I also changed the incorrect schottky diodes to the correct zenner diodes as well.
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
O.K., now I see the schematic. IMHO it's a bad practice to change pictures in previous posts by reusing the image name in new posts and replacing the content. It's like manipulating history because the edit isn't reflected by a change notice in the posts itself.

You are right that the 1N4007 are useless for protection. I also don't understand the purpose of the snubber circuit. The MOSFET driver should be primarly protected by bypass capacitors and power zener respectively transsorb diode limiting the supply voltage to the allowed range. Both TC4422 VCC pins are internally connected, they should be used both and connected externally, too.
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
Both TC4422 VCC pins are internally connected, they should be used both and connected externally, too.
What difference does this make? Is it easy to explain or something I should just take your word for? Does that also mean the GND pins need to be externally connected?

The MOSFET driver should be primarly protected by bypass capacitors and power zener respectively transsorb diode limiting the supply voltage to the allowed range.
I understand that the mosfet is protected by the series zeners, but apparently that does nothing for the TC4422 as it keeps frying on me. So clearly I need another layer or protection in here specifically for the TC4422.....hence the snubber circuit I am trying to figure out.

When I connect this circuit up to a spent car battery, that is probably lucky to output 1A, the TC4422 appears to be OK as the circuit just keeps running. But as soon as I hook the circuit up to a good car battery that puts some where between 1 and 4 amps (4 amp fuse in the 12V supply rail that is not depicted in the schemtic) the TC4422 dies pretty quickly. Perhaps with the crappy battery the TC4422 is still dying but much more slowly.
 
Last edited:

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
Does that also mean the GND pins need to be externally connected?
The datasheet tells to connect all GND pins.
I understand that the mosfet is protected by the series zeners, but apparently that does nothing for the TC4422 as it keeps frying on me.
Of course not because absolute maximum rating is 20V.

What most likely happens is that the wiring inductance between battery and output transformer generates a kickback voltage that exceeds the said 20 V. Massive bypassing and Z-diode protection with suitable voltage level should be able to protect driver and MOSFET gate.
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
The datasheet tells to connect all GND pins.

Of course not because absolute maximum rating is 20V.

What most likely happens is that the wiring inductance between battery and output transformer generates a kickback voltage that exceeds the said 20 V. Massive bypassing and Z-diode protection with suitable voltage level should be able to protect driver and MOSFET gate.

Ohhhhhh....of course!
The penny has finally dropped FvM so thanks.

I knew that kick back was killing the TC4422, but what I was forgetting was that the Jaycar jacobs ladder kit documentation states that they have only limited the kick back voltage to within the limits if the Darlington transistor they used. And that is something like 200V from memory - well in excess of the TC4422 limits.
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
I knew that kick back was killing the TC4422, but what I was forgetting was that the Jaycar jacobs ladder kit documentation states that they have only limited the kick back voltage to within the limits if the Darlington transistor they used. And that is something like 200V from memory - well in excess of the TC4422 limits.
In a world of ideal designs, there's no wiring inductance, only the transformer primary inductance. But in practice there's wiring inductance. If it's only 0.1 µH versus 1 µH transformer inductance, you get 10 % of the inductive kickback at your 12V supply, but without any limiting means.
 

boylesg

Advanced Member level 4
Joined
Jul 15, 2012
Messages
1,012
Helped
5
Reputation
10
Reaction score
5
Trophy points
1,318
Location
Epping, Victoria, Australia
Activity points
11,638
In a world of ideal designs, there's no wiring inductance, only the transformer primary inductance. But in practice there's wiring inductance. If it's only 0.1 µH versus 1 µH transformer inductance, you get 10 % of the inductive kickback at your 12V supply, but without any limiting means.

The zener diodes disipate kick back voltage in excess of 300V (4 x 75V). So that leaves up to 300V swirling around the circuit and into the TC4422.

So, on reflection for a day, simply adding a parallel (to the TC4422) zener diode 1N4745 to the circuit should be sufficient to disipate kickback voltage in excess of 16V which is well within the limits of TC4422 (20V max).

That is assuming that the kick back current is small enough such that V * I does not exceed 1W for the zener.

Your opinion on this modification to my circuit?
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,914
Helped
14,335
Reputation
28,933
Reaction score
13,080
Trophy points
1,393
Location
Bochum, Germany
Activity points
282,243
It's at least a step in the right direction. The kick back current can be as high as the inductor current. That's no problem for the z-diode if the curent pulse is short.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top