Bride of Zen with cascode and current source.

Status
Not open for further replies.
Jump to Latest
#21
:idea:
As I was laying in bed I realised the solution tho my posted question (I think!). The answer must be generated iteratively. In other words, if we assume that Vgs has a value of 4V we can look in the chart over IV-characteristics for the MOSFET and see which Ids this givs us (using the load line). Then using kirchoff we get the voltage drop over Rs. Now we can see if this voltage together with Vgs is close to the voltage at the gate of the MOSFET. If so, then we have found the working point, and if not we guess a new value of Vgs and so on and so forth. Of course this only works if we have a accurate chart of the MOSFETs IV-characteristic, which we don't, which means that we have to create the circuit and see what we get, alternatively measure the MOSFET.... :bigeyes:

I had some trouble falling to sleep!
 
#22
There is a reason they call this DIY.
Click to expand...

Hi Nelson,

normally I'd do this at once, using a cascode stage and compare it to the non-cascoded simply by removin the MOSFET. But in my case I would switch to another topology with a folded cascode. Unfortunately in this case nearly every voltage must referenced to another point. Because I'm using a PCB I have to do 2 PCBs. With a power supply and a relay attenuator my PCB is about 12 x 16 inch. Only one PCB will cost me about $400. BTW did anyone said DIY is cheap?:bawling:

So I hope you can help me to make my decision a little bit easier.
 
#23
If so, then we have found the working point, and if not we guess a new value of Vgs and so on and so forth. Of course this only works if we have a accurate chart of the MOSFETs IV-characteristic, which we don't, which means that we have to create the circuit and see what we get, alternatively measure the MOSFET....
Click to expand...

You got it. Designing with bjts is easier, because the have a very sharp Vbe = 0,6V when biased. MOSFET parameter variation is higher. Matching MOSFETs is one solution another is to make the current flown through the MOSFET as much independant as possible wich means that you have quite a large voltage from the source to ground. Or - last solution - just use a trimpot to adjust Vgs.
 
#24
Stäppvargen, before you have come too much ahead, investigate which transistor you really can get! Don't use a SOT23 if you don't have special reasons. 2N7000 is the same type in TO92. Why don't you try BS170 of PMBF170 (SOT23)?

I can't see that 2N7002 is better than 2N7000. It's the same chip but the SOT23 can't handle so much power.
 
#25
bocka wrote:
Designing with bjts is easier, because the have a very sharp Vbe = 0,6V when biased.
Click to expand...
Thanx, point taken. All the possible solutions are giving me a headache! Perhaps I will just buy a number of mosfets/jfets/bjts and test them in different configurations. I already have a nuber of IRF610 and was planning to buy some 2sk170. Do you have any recommendations of bjt candidates and possible schematics for them?

peranders wrote:
Stäppvargen, before you have come too much ahead, investigate which transistor you really can get! Don't use a SOT23 if you don't have special reasons. 2N7000 is the same type in TO92. Why don't you try BS170 of PMBF170 (SOT23)?
Click to expand...
I'm sorry I wasn't very clear. I have no plans of actually using the 2N7002, but it was only chosen because it was the only smallsignal-MOSFET spice model available to me. You know what my choices mainly consist of, namely components from ELFA (www.elfa.se/se) and RS-Components (www.rs-components.se). Any thoughts?
 
Status
Not open for further replies.
Top Bottom