0

2907 TUBE DRIVER [Replace Q3, P channel MOSFET]
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

Re: 2907 TUBE DRIVER

by Russell 27 on Sat Oct 12, 2013 8:28 am

Most of our discussion has been about the right current and voltage in driving the display, this is of course important. But I've noticed another "physical" aspect. When I bought the kit I also purchased an extra tube and pcb header. I've used both of the tubes in my clock. Both are the 19 pin version and both have decent display quality, driven by the kit circuit. I noticed one tube's PM indicator dot is a little dimmer than on the other. Upon examination of the tubes, I discovered the filament was almost perfectly parallel to the grid, in one. The other, the filament was at a slight angle, closer on one end of the segment grid span than the other end. Guess which tube showed the the dim PM indicator dot? Actually, the parallel filament tube, almost seems brighter across the entire grid, at the same brightness setting. Driving the tube at a higher grid/ segment voltage probably helps even things out, or the eye may just not notice this as much in a brighter display. At a lower driving voltage this can become more noticeable. It seems that this physical aspect of the filament, does contribute to the overall gradient, and then driving voltage is more important.

Russell
Russell 27
 
Posts: 240
Joined: Thu Sep 12, 2013 3:59 pm

Re: 2907 TUBE DRIVER

by jarchie on Sun Oct 13, 2013 2:54 pm

Russell 27 wrote:I noticed one tube's PM indicator dot is a little dimmer than on the other...It seems that this physical aspect of the filament, does contribute to the overall gradient, and then driving voltage is more important.

Many months ago, I spent a day playing with a half-dozen tubes at different AC and DC voltages across the filament and different grid/anode voltages. From that testing, I inferred several major factors affecting display evenness, summarized below. Hope these observations are useful!

The voltage gradient issue: In the Adafruit design, the negative side of the filament is on the left and the positive side is on the right. So the tube will be brightest, on average, on the left side, because the voltage between the anode and cathode is greatest on the left. When the gradient is the predominant issue, increasing filament voltage makes the gradient worse, while increasing the grid/anode voltage decreases the gradient. Strangely, driving the filament at ~3.5v DC makes the gradient almost unnoticeable, even at low brightness (low grid/anode voltage), which is how the filament is driven in the Adafruit design. EDIT: The Adafruit design actually runs the tube at well below 3.5 volts.

The dim digit issue: With insufficient voltage/current across the filament, the rightmost digit and (less commonly) the indicator appear significantly dimmer than the other digits. I believe this issue is caused by insufficient heating of the filament near the filament attachments, which conduct heat away from the ends. If DC filament voltage is increased, the dim digits brighten, and the display becomes even. If the grid/anode voltage is increased, the rightmost digit becomes even more dim with respect to the other digits. I suspect that, with a higher grid/anode voltage, more electrons are "robbed" from the filament by the segments, thereby reducing filament current/temperature at the rightmost end.

Intrasegment variation: At low brightness, individual display segments exhibit slightly uneven brightness. Increasing the grid/anode voltage evens out the individual segments, but also makes the display brighter. But if the grid/anode voltage is kept high, the display can also be dimmed by applying a PWM signal to the grid/anode, providing low brightness with more consistent segment illumination.

Tube variation: Some tubes have a natural gradient, while others seem to have one digit that is slightly dimmer than the others. These issues, when due to tube variation, are much less noticeable than the display gradient or dim digit issue mentioned above. Changing the filament or grid/anode voltage has less of an effect on these issues, but driving the display closer to specifications seems to help a bit.

Also, the severity of the display gradient issue and the dim digit issue differ slightly between tubes. (For example, some tubes exhibit the dim digit issue at higher filament voltages than other tubes.)
--John <www.jarchie.com/email>

jarchie
 
Posts: 608
Joined: Sun Jun 24, 2012 2:16 pm
Location: Santa Cruz, California, United States

Please be positive and constructive with your questions and comments.