ICE TUBE Q3 TEST: MOSFET or PNP TRANSISTOR Moderators: adafruit_support_bill, adafruit

[jarchie]

I'm not arguing anything, the evidence is right in the meter shots, earlier in the post. No noise.

The meter shots do a good job of illustrating the electrical characteristics of the transistors, but without examining a clock with a flaky segment display problem, how can you claim that those clocks have no noise on the output of the Q3? I was fortunate enough to actually examine a clock exhibiting a flaky segment display problem, and the scope pictures show noise quite clearly. My measurements confirm those of Frank_tt who also observed noise.

[phild13]

Frank_tt had issues with the kit supplied ZVP3306A which we know does not fully turn on and depending upon part variation, produce a flaky display on some clocks (we are just seeing a subset of kits sold posted here). I would expect it to produce lots of noise since it does not fully turn on and tends to oscillate a bit because of that. Oscillation appears to be evident on the scope measurements he included in the post. Following the same thread, the solution Frank_tt posted was to first turn around Q3 so it was installed backwards which caused the display issue he was having to go away but killed the battery. Then he ordered a ZVP2110A and put it in the correct way (matching boards silkscreen outline) and that also eliminated the display issues and did not kill the battery.

From my perspective, both you and Russell are looking at Q3 in a different way. As you noted in the past your using a scope and looking for/at oscillations on a component that appears to cause voltage passing through the component to vary which I would think could if the oscillation dips low enough for a long enough period cause the VFD chip to be affected. Scope photos

Russell is looking for/at voltage input vs voltage output on a component with a standard meter to demonstrate how well a particular component passes voltage when it's base or gate is turned on via the clock circuit. The meter is a slow response averaging device and can only really see the average voltage at the point of measurement and so will not show any oscillations as a voltage change unless the time period of the oscillation is great enough for the meter to see. As the time period of the oscillations is short the meter will not see them. Application of the meter may also kill any oscillation present, thereby essentially eliminating it during the period of measurement.

The end result of both methods of testing/measurement is that the better more efficient the component is at passing voltage from input to output the less affect any voltage variations or oscillations will have on the function of the circuit downstream of the component. Your scope photos show that and Russell's voltage measurements also show that.

Personally, I like the ZVP2110A as a one part replacement as that seems to work for anyone who has tried it and their clock was not damaged in some other way. I also like the transistor/base resistor combo as it also works very well. Each component has advantages and disadvantages. A FET easy to damage with static during handling before install but costs less to implement and the transistor requires soldering base resistor to base lead which increases risk heat damage and costs a bit more to implement for the needed base resistor.

[Russell 27]

In retrospect, I regret making this post. I should be able to express my ideas here, especially the ones I tested and verified. When I can't ask a question on my own post, my information is interpreted another way, and something else is recommended instead; The forum is here to try new things, not in my case I guess. I hand it over to the all knowing, ICE TUBE MASTER.

[jarchie]

I'm truly sorry that this is how you feel. I honestly believe that the testing you did to compare the various candidates for Q3 is extremely helpful. I also agree that low filament voltage does cause some types of display problems.

We might not agree on the cause of flaky segment display problems, but that's only one point of disagreement. We seem to agree that the ZVP3306A does not turn on fully and that using a better part will solve the issue. The results you've posted clearly show that there are, in fact, better parts to replace Q3.

[phild13]

Based on the results of Russells testing of transistors/FET's and his testing of SMD trimmer caps for C8, C9 in another post, I have gone back and made some mods in addition some other minor mods to the transistor main board I designed so that I can implement both the transistor/base resistor and the trimmers. I'm going to wait until I have the trimmers in hand as I might just make SMD pads instead of through hole.

At any rate I think I will use the 2907 and a 470 ohm and will also pick up a few BS250 while I am at it. As for the caps I think I will pick up a couple of the white ones Russell used in his mod posted photos and a couple smaller ones I was thinking about.

I would not have even bothered with any of these mods had efforts by others not been made to do testing, post the results, and discuss the results which got my interest up enough to pursue some of these mods.

I don't care about certain aspects or if no one agrees on everything. Since I'm from the tube era, the discussions have caused me to get back into something I left all to long ago and to study up on things that never were any concern to me.

Heck, all I was intending to do was to put together a kit for relaxation.

So don't anyone feel their time or effort was wasted or the discussions and different points of view are not helpful, they are and I'm sure I'm not the only one who has gained some greater insight into the clock and electronics in general.

[Barry914]

If all I wanted was a clock that kept time I would have gone out and bought one. Troubleshooting, experimenting, making mistakes and hearing what others have to say are all part of the fun. We are doing this for fun, aren't we?

[phild13]

I have had more fun with an icetube than a store bought clock.by far and enjoy others comments and input. Throughout I have learned much about the design of the clock, how it all fits together and how to improve upon it.