Hi, i have designed a simple linear regulator with a voltare reference and a op amp for error amplifier. As pass device i've used a big IGBT (BSM 300GA 160D) and i've noticed that if i put in the input of regulator a 100 Hertz sinusoidal tension when the regulator start to "clamp" i have not a clamped sinuus but there is a spike at the begin then the regulator clamp.
If e put as pass device a mos(irf 450) the sinus is totally well clamped. Why? What is the problem? The gate of igbt and mos is drived direcly by the op amp. I've used also a ICL7667 diver (i thougth that igbt need more current) but rrthe resul is the same with IGBT spike at hte begin od sinus. What is the error? Thanks
Search on "stored charge", "base charge" and "the stored charge model" for bipolar transistors. The base-emitter junction of a bjt is a voltage dependent capacitor (varactor), as is the gate-source junction of a MOSFET but igbt's are slower to clear out the stored charge
Search on "stored charge", "base charge" and "the stored charge model" for bipolar transistors. The base-emitter junction of a bjt is a voltage dependent capacitor (varactor), as is the gate-source junction of a MOSFET but igbt's are slower to clear out the stored charge
Do not let the Slowman confuse you with gate charge talk. At 100 Hz this will very likely be insignificant. The reason you see the spike is more likely because the IGBT needs some voltage between collector and emittor to 'open', say 'to start conductiing', while the MOSFET acts much like a low value resistor when 'on'. Look at the IGBT curve here:
Do not let the Slowman confuse you with gate charge talk. At 100 Hz this will very likely be insignificant. The reason you see the spike is more likely because the IGBT needs some voltage between collector and emittor to 'open', say 'to start conductiing', while the MOSFET acts much like a low value resistor when 'on'. Look at the IGBT curve here:
I don't know what you are using to monitor this, I can only assume a scope? In any case, older IGBT's had slow switching issues in the BT output much like slow switching transistor would be. Also, some IGBT's have an internal series R to the gate. with the high capacitance found on these devices it's possible this is what you're seeing that propagating the delay..
look up the spec's on the switching time and check for internal R to the gate etc..
It does happen to be a popular way of modelling turn-on and turn-off in biploar transistors, but pantel prefers to be confused in some other way, which he hasn't bothered to tell us about.
Because?
In other words, the igbt is taking time to turn on, but pantel isn't interested in pointing out a way to quantify this delay.
which is precisely the same web-site that I used in my attempt to confuse maxi.
Yes a check with the scope on the emitter. The device is used as pass power in a linear regulator. When the sinus(100 hz) is higher of reference igbt should off,but igbt off with delay and i've a part of sinus that pass. With a 2N3055 or a IRF450 the sinus is immediatly clamped. Is there a circuitation can switch off the igbt faster.
"Jamie" ha scritto nel messaggio news:X%JMl.3764$ snipped-for-privacy@newsfe19.iad...
command igbt ad emitter. I havent' condenser..........i must clamp a sinevave,100 hz is full wave rect. Thnks "Jan Panteltje" ha scritto nel messaggio news:gu11ib$nv4$ snipped-for-privacy@news.albasani.net...
On a sunny day (Fri, 08 May 2009 11:29:42 GMT) it happened "maxi" wrote in :
OK, so if I look at hose waveforms, are you having the IGBT in series or to ground?
Some versions of IGBTs could have 'latchup', where they would keep conducting, in a series configuration that could explain why it stays on until the voltage reaches top, and then reverses?
If you 'clamp' to ground then I would like to see the circuit (in any case), else it will be a lot of guessing.
On a sunny day (Fri, 08 May 2009 14:02:01 GMT) it happened "maxi" wrote in :
Interesting, not sure I would do it that way, maybe I do not understand it?
AC in | ---1M --- + | | opamp -------------- series reg | -- - | | | | | ---------------------------------- | | 5V R 10k | | ---------------------------------------------- clipped AC out | 75k | -AC
I think you are not actually measuring the output voltage in the correct way !
Look at the emittor connected to the - of the opamp. On the + of the opamp is some DC voltage with the negative AC side superimposed on it. If we leave that AC coupling out for a moment, then you have created a current source! It will try to maintain a current 5 / 10000 = 1/2 mA in the 75 k resistor.
If we now add the capacitor of 100 nF from -AC to the + input from the opamp, then we indeed modulate that current, but it does not know about absolute voltage level!
If you wanted to clamp, or rather _limit_ the positive going voltage, then you need to measure that. For example like this:
+AC | ----- | | | | | 1k | | | |
|----------- + |
15V | opamp -------------- series reg
- | ------- - | | | | | | | zener R1 10k pot | | | | | | ---------)-----------------------------------| | | A | | clipped AC | | | -----------------------------------| B | -AC If R1 =1 M, and the pot is 10 k, then if point A is more positive then point B, so that the - opamp input becomes more positive then the + opamp input, then the series regulator (whatever it is) will stop conducting. You still need the protection diodes to protect the opamp inputs, but no capacitor.
So for 10 k and 100 V positive clipping, and 400 V pp betwen +AC and -AC, you have about 500V over R1, and 5 V over the 10k pot, that makes R1 about 1M.
The AC input is +AT Lnea e - AT linea,so i connect the scope probe to Emitter and - AT linea and i have the clamp. The 10K trimmer is for regulating the treshold of clamping. The problem is that with the Mos the clamp is ok,but with the IGBT the clamp is as the picture posted.
It seems to not clamp at the start of the sinus and after few microsecond the igbt start to clamp(not conduction).....but really i dont't understand why. I'd like the igbt because sometimes the AC sinus can arrive at 1200 volt ad also is a 1600 volt 300 ampere....whre i'll sink 10 ampere and clamp all at the max of 350 volt "Jan Panteltje" ha scritto nel messaggio news:gu1m5t$m3o$ snipped-for-privacy@news.albasani.net...
On a sunny day (Fri, 08 May 2009 17:04:15 GMT) it happened "maxi" wrote in :
I thought I explained that, here is the spice run of your circuit: ftp://panteltje.com/pub/italian_limiter.gif Does not look good even with a MOSFET, If I add the diodes it does not even run at all: ftp://panteltje.com/pub/italian_limiter_with_diodes.gif Did I make an error in the diagram?
And here is the spice run of mine with the same MOSFET and opamp: ftp://panteltje.com/pub/limiter.gif Nice and flat. No diodes needed, no capacitors needed.
i've used your circuit but isf i set for example the sinus at 400 vol ac or
600 the output(gnd on ground ac generator and drain of mos) is zero volt! why? "Jan Panteltje" ha scritto nel messaggio news:gu1rus$u1f$ snipped-for-privacy@news.albasani.net...
On a sunny day (Sat, 09 May 2009 11:22:01 GMT) it happened "maxi" wrote in :
Not according to spice, here is it with 600V AC: ftp://panteltje.com/pub/limiter_600v.gif
When you say 'drain', the drain however is connected directly to the AC! So did you perhaps mean to say 'source' of the MOSFET? Did you wire it up correctly?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.