LTSpice results not what I expected with a 2N3055 transistor

Thread Starter

TrackerFX

Joined May 18, 2020
13
I'm building a bench power supply and would like to switch between several outputs. I have a combination volt/ammeter that I would like to use to monitor the selected output while still being able to use the other outputs. The idea was to use a pair of 2N3055 transitors in parallel (one with a logical inverter on the input) to switch the outputs between direct access to ground and access to ground through the ammeter. The problem came when I made the circuit in LTSpice and the results were not what I expected. In order to troubleshoot the circuit I reduced it to the bare minimum that still returns the same results and this is what I ended up with (the graph shows the voltage and amperage at the collector):
Screenshot (2).png
As you can see, the load is only dropping ~3.5V while the transistor is dropping the rest. I chose the load size to simulate the max amperage that I would allow through this circuit. The datasheet for the 2N3055 says VCEO=60VDC and IC=15A. Is there a mistake with the Spice model, or am I missing something here?
 

ronsimpson

Joined Oct 7, 2019
680
Voltage across R1 is 5-0.7=4.3V. The current in the Base of Q1 is 43mA. Not much to switch 10A.
At 10A you only get a gain of 15. typical
2.5V on 2.4 ohms is about 1.5A in the collector.
At 1.5A in the collector the gain is about 90 on a average part but could be much lower.
You are only getting a gain of 35.
Data sheet "at 4A in the collector the gain will be anywhere from 20 to 70". (not a direct quote)
"at 10A the gain is larger or equal to 5"
--edited--
Short answer is you need much more Base current.
Have you though about using a MOSFET?
 

Papabravo

Joined Feb 24, 2006
13,728
Show us the 2N3055 model you are using. Since transistor parameters are not well controlled you should expect the numbers to be a bit...surprising.
BTW 20V of Vce says you are just beginning to turn the transistor on. You need an order of magnitude more base current -- this part is a beast.
Try using a behavioral current source and sweep the current from 0 to 1 Amp and back.
 

crutschow

Joined Mar 14, 2008
24,968
The 2N3055 may have a max rating of 15A, but it has very poor switching characteristics at currents even half that.

Below is an LTspice simulation of the 2N3055 model I have.

As you can see, the current-gain (red trace) drops to below 10 at 8A output with a 1V (arbitrary supply voltage) drop across the collector-emitter.
Horizontal scale is base current.

1589867502548.png

I suggest using MOSFETs.
They require no current, just voltage to turn on, and the low on-resistance devices can have only small fraction of a volt drop at 10A.
 
Last edited:

Papabravo

Joined Feb 24, 2006
13,728
The 2N3055 may have a max rating of 15A, but it has very poor switching characteristics at currents even half that.

Below is an LTspice simulation of the 2N3055 model I have.

As you can see, the current-gain (red trace) drops to below 10 at 8A output with a 1V (arbitrary supply voltage) drop across the collector-emitter.
Horizontal scale is base current.

View attachment 207560

I suggest using MOSFETs.
They require no current, just voltage to turn on, and the low on-resistance devices can have only small fraction of a volt drop at 10A.
You must have a great deal of time on your hands -- grin
 

Thread Starter

TrackerFX

Joined May 18, 2020
13
Since the internal resistance of my supply is near zero I had not added resistance I tried adding low value series resistance to the sources but that didn't change the results by much. Switching to the MOSFET seems to do the trick, though! Thanks for the help!
 
Sharing/switching an ammeter electronically has always proven problematic. It is easier to provide a current measurement shunt for each source, then use a simple multiplexer to switch the meter that reads the voltage dropped across the selected shunt. I've successfully removed the shunt from an ammeter module, then remotely located it to minimize the high-current path. Perhaps you can do the same but with duplicated shunts and an added multiplexer.
 

MIS42N

Joined Jan 25, 2013
3
I am amazed the 2N3055 is still used. I bought one over 50 years ago to use as a series regulator in a 50 watt valve amplifier, a pair of 6CA7/EL34 in ultra linear configuration. The '3055 dropped the unregulated 450V to regulated 400V, mounted on a heat sink on insulated standoffs. Without feedback this gave less than 1% THD. Later I built a transistor stereo amp with pairs of '3055 as output transistors. I think the circuit was a Playmaster series from the magazine Electronics Australia. Completely off topic, I was feeling nostalgic.
 
Seems like a lot of work, plus requiring the purchase of additional identical shunts, when a couple MOSFETs per channel should work fine for that purpose.
Could be less work, as a shunt is nothing more than a low value resistor - readily available from online distributors. The MOSFET approach requires multiple high-current transistors, careful make-before-break sequencing, and will result in a change of output voltage when the switch is made - unless a "fake" shunt is switched in using matched MOSFETS. That and the supply output impedance is bound to be higher / the regulation poorer due to the added MOSFET resistance. FWIW if a gain stage is added after the proposed mux and prior to the modified ammeter it's possible to use external shunts that are even lower in resistance than the original.
 

dgwheeler

Joined Oct 29, 2019
1
I'm building a bench power supply and would like to switch between several outputs. I have a combination volt/ammeter that I would like to use to monitor the selected output while still being able to use the other outputs. The idea was to use a pair of 2N3055 transitors in parallel (one with a logical inverter on the input) to switch the outputs between direct access to ground and access to ground through the ammeter. The problem came when I made the circuit in LTSpice and the results were not what I expected. In order to troubleshoot the circuit I reduced it to the bare minimum that still returns the same results and this is what I ended up with (the graph shows the voltage and amperage at the collector):
View attachment 207556
As you can see, the load is only dropping ~3.5V while the transistor is dropping the rest. I chose the load size to simulate the max amperage that I would allow through this circuit. The datasheet for the 2N3055 says VCEO=60VDC and IC=15A. Is there a mistake with the Spice model, or am I missing something here?
I use 6055 darlington transistors. Good for 8 amps each and have a gain of 100@8A. Use a .01 resistor in the emitter of each to force equal current sharing.
My 2 cents worth!
 
Top