Zero to -40v Fast Switch Driver

[grizzly1]

I need to create a fast swtich driver to switch fro zero to -40 volts - I believe I can use a jfet or mosfet but do not have a lot of experience with them so I am diving into unknown territory a bit.
I am switching a GaN RF device and it draws little to no current -- < 1 mA. I need to command with a TTL singnal which I can probably use some simple gate as a buffer . I need to switch in <100 nS, the faster the better but that is a goal at this point. Any suggestions on devices and a basic circuit to get me started would be highly appreciated --

Thanks in advance

Griz

 

[grizzly1]

I also have +5 available -

 

[alec_t]

Is your TTL signal floating or ground-referenced? In other words, is logic 0 at zero volts, or could it be at -40V?

 

[grizzly1]

Standard TTL logic 0v - 5v referenced to ground

 

[jpanhalt]

Last Winter I played with a negative voltage switch. I don't know how fast it can switch, but here is the circuit and simulation.

View attachment 66995

John

View attachment 66996
View attachment 66997

 

[ccurtis]

I am switching a GaN RF device and it draws little to no current -- < 1 mA.

That's interesting. What meaningful output power can you get from a RF device generally used as an RF power output stage when it draws so little current? If you are switching the bias (which is low current), do you know that you are not supposed to switch the gate (bias) to off with drain voltage applied? Generally you switch the RF input on and off, if anything. Just curious.

 

[grizzly1]

John -- thanks I will investigate your schematic further it at least gives me a starting point, I am having trouble reading the schematic what is the FET ? Looks like a P channel enhanced FET ? Can you post the part # please ? I also assume you are taking the output voltage off the drain? I am unable to open either the .asc or .sch files

ccurtis - the RF device is a GaN SPDT switch manufactured by Tri-Quint there are 2 versions one DC-6 GHz and one 2-18 GHz the low F one handles 40 W and the high band around 20 W. On their website --

 

[jpanhalt]

The .asc file opens with LT Spice (and probably other spice simulators; I haven't tried). The .sch file opens with EAGLE. For that, you will need to download and install the free version.

I have attached a png of the schematic from the Eagle file.

John

View attachment 67015

 

[grizzly1]

Thanks John that gets me started - I get the idea with this schematic -- appreciate your time

Griz

 

[alec_t]

Not sure if the 2N3906 in the above circuit will be quite fast enough; but no doubt a faster transistor could replace it.
Here's a variation, using a 'high-side' switch (as we don't know quite how your system is configured). The 1 mA load (GaN device) is represented by the 39k resistor. The inverter (an arbitrary choice) is just to demonstrate driving the switch from a logic device. The top end of the load is switched from +5V to -40V in the circuit shown, but changing the negative rail to -35V would result in a 40V swing instead of the present 45V.
As simulated, switch-on time is ~ 40nS, but switch-off time is ~ 150nS.
View attachment 67020

 

[grizzly1]

Alec - I think the problem with this is that the control input to the GaN switch can not go positive more than a couple tenths of a volt I think when the transistor is on the top of the resistor will be at a positive 5 volts - The GaN switch logic input is 0v and -40 volts the negative turns the switch off and the 0 volts switches it on

 

[grizzly1]

Here is a link to the switch

https://www.triquint.com/products/p/TGS2351

 

[alec_t]

Ok. This version switches 0V to -40V with similar on/off rates to the post #10 circuit. Resistance values are kept low-ish to reduce switching times, so paralleled gates share the drive current.
**broken link removed**

 

[grizzly1]

Sorry Alec but the link does not work -- it gives me an error message

 

[ronv]

There is a link to drivers in the data sheet.

https://www.triquint.com/products/p/TGS2351

 

[grizzly1]

Thanks Ron -- what is really interesting is that this must be a new link -- When I first started with this switch project, this was not available -- I had worked with one of thier engineers and he did not have this when these devices first came out -- thanks for pointing it out to me -- I will explore this further

Thanks all for the help it is highly appreciated

Griz

 

[alec_t]

Better luck this time. Previous link must have got corrupted.
View attachment 67030

 

[grizzly1]

Got it this time alec -- thanks -- I think I have a plan now based on all the inputs the ap note that Ron pointed out to me was very helpful as well as everyone else's inputs -- First time I have posted here and found everyone freindly and willing to help -- thanks all

Griz