Analysis of 81936 A9 Power Distribution Board - Orion 565 and Orion 566 OII

October 17 , 2007

This board supplies 5 vdc and the -20.5 vdc for the 565 display The same board is used in both 565 and 566 but the -20.5 vdc portion is not needed for the OII display

I did this analysis as I have had two of these boards fail and am aware of several others having failed and wanted to understand the root cause. Generally the failure is C6 the 470 ufd 16 volt Illinois Capacitor 477CKR or CKS 016M .

I have measured the case temperature of the 477CKR016M C6 to be 60 deg C in my Orion even with the case open and an ambient of 25 deg C. With the case closed I have observed an ambient in the area of the 81936 board of 46 deg C after just two hours so I expect the case temp of C6 could be very close to 80 deg C under many conditions.

I did a Spice analysis of the entire board using LT Spice from Linear Technologies who supply the LT1054 used in this board. I have incorporated the latest resistor changes as of when my 566 was shipped 5 Dec 2006.

I used a source of 13.8 volts with a series resistance of 0.2 ohms. Through indirect means , (measuring the DCR of L2 , the 22 uh choke in the output and the runner to the 5 volt pins , and then measuring the DC drop across the path I determined the average current was about 1.5 amperes.)

With a simulated output of 1.52 amperes at 5.055 volts I found the rms currents in C6, C7, C9 as follows:

C6 1530 ma rms , V avg 5.714 volts

C7 248 ma rms , V avg 5.08 volts

C9 760 ma rms , V avg 13.6 volts

If we look at the original 477CKR016M -or 477CKS016M (they are now a merged line) it is a 2000 hour cap at 85 deg C and rated at 470 ma ripple at 120 hz. Even assuming 60 deg C case and 6 volts - using their data sheet we get a ripple rating of a 1.36 multiplier for 25 khz ripple and a 1.5 multiplier of ripple current for 60 deg C case . This is 470 X 1.36 X 1.5 or 959 ma. So for C6 we are already running the capacitor well above ripple ratings even at 60 deg C case. Then using a more likely 10 deg C delta (85-75) will give a temp-life multiplier of about 2 and voltage of 6 is a life multiplier of 2 (Ill Cap says if you are below 1/2 the rated voltage the 2X applies but less than 1/2X buys very little) so 2000 X 2 X 2 is 8000 hours or about a year 24/7 but who knows how much it is lessened by running ripple WAY above rating - so failure in a year or so is not unbelievable at all- depends on how much it is run each day.

I searched for another capacitor and talked with Illinois Capacitor twice to be sure I understood their data sheets (and found an error in their data sheets) for the 477KXM016M thru 035M capacitors which looked like a possible candidate for a replacement . They have since changed the detailed data sheets on these capacitors on the web and will change them in the new catalog).

Illinois Cap suggested I look at the KXM series since it is a 5000 hour capacitor.

I purchased some 477KXM035M capacitors and installed them in place of C6, C7 and C9. The temperature rise and board temperatures an ambient temperatures are shown below.

With the case open I get the following :

Ambient (air) temp in vicinity of the board 28.4 C Printed circuit board temp at base of the two capacitors C6,C9 46.7 C
C6 case 47 C (1530ma)
C9 Case 44.7 C (760ma)
C7 case 40.2 C (248ma)

Then I slid the Orion top and bottom covers on and left a thermocouple about 1 inch from the board to measure ambient . I ran it for 2 hours transmitting at 35 wpm for 38 seconds - 10 seconds receive on 14 mhz into a dummy load at 100 watts indicated.

Under those conditions I get the following (cap case temps measured as fast as I could get them after sliding the lids off - and power still on while measuring - but not transmitting)

Ambient (air) temp in vicinity of the board 45.5 C.

FWIW -the ambient went as follows during the closed heating cycle. (power was off while replacing covers - then turned back on)

Time Temperature
3:20 28.4 C
3:34 32.0 C
3:54 37.2 C
4:21 41.4 C
4:46 43.8 C
5:23 44.8 C
5:47 45.5 C

Printed circuit board temp at base of the two capacitors C6,C9 55 C
C6 case 56 C (1530ma)
C9 Case 53 C (760ma)
C7 case 46.2 C (248ma)

Then I re-measured again as it was cooling (obviously the ambient temp was changing as I was measuring the above - and you never know how long to wait for the next digit to change on the dvm)

Printed circuit board temp at base of the two capacitors C6,C9 51.8 C
C6 case 50.3 C (1530ma)
C9 Case 47.0 C (760ma)
C7 case 42.8 C (248ma)

There is significant heating of the capacitor from the 1.8 watt dissipation in Q2 transistor right by these capacitors as well as the internal heating of the capacitor and other losses on the board. Total dissipation on the board is 3.5 watts based on an input of 11.2 watts for an output of 7.69 watts.

I might point out that if my assumed 1.5 amps out is low , the 1530 ma will very closely just be ratioed up or down by the actual current divided by 1.53 .

The revised data sheets dated 10/15/2007 on the web site for the 477KXM025M and 035M are as follows.

477KXM025M Ripple current at 100khz and 105 deg C is 920 ma.

Ripple current multipliers are 0.99 for 25 khz and 1.65 for 60 deg c case temp , yielding 1503 ma for this application. This would be borderline for the C6 location, OK for C7 and C9.

477KXM035M Ripple current at 100 khz and 105 deg c is 1120 ma.

Ripple current multipliers are 0.99 for 25 khz and 1.65 for 60 deg c case temp , yielding 1829 ma for this application. This would be even more satisfactory for the C6 location.

I might also note these KXM series capacitors are rated for 5000 hours of life with 100 percent voltage and 100 percent ripple current applied. (Ill Cap says for the approximately 6 volts across the capacitor a life multiplier of 2 can be used and for 60 deg C a life multiplier of approximately 22.7 can be used. (This comes from the approximation of double life for each ten deg c below 105 C - or 45 C delta.) So the expected life of the 477KXM035M would be 5000 x 2 x 22.7 or 227000 hours. I was told that life is determined mainly by voltage and case temp as long as ripple current is less than ratings.

Finally , in conversation with Illinois Cap , they also added they restrict the case temperature to 10 deg C above ambient to maintain the rated life , irrespective of ambient . In other words they do not like to see enough dissipation in the capacitor to get more than 10 deg C rise in case temp above "ambient" . I could never get Ill Cap to just talk about case temperature.

For all the above reasons , I would suggest the 477KXM035 cap at C6 and the 477KXM025M could be used at C7, C9 if the cost differential was enough to overcome the handling issue of two capacitors in that board.

I was uncomfortable with the Ill Cap case temp /ambient discussions so looked for another candidate. Panasonic came highly recommended by a friend who worked in a Motorola group building 24/7 computer systems and they had made a change to Panasonic after some similar capacitor life/ripple current issues.

A Panasonic candidate is from the FM series . Their EEUFM1E471 is a 470 ufd 25volt 105 deg C capacitor rated at 1790 ma ripple current at 100khz. The multipliers for ripple current are 0.99 for 25 khz and they do not specify a temperature multiplier for case temperature versus ripple current but do use the case temp multiplier for life calculations. This cap is a 4000 hour rated at 105deg C and 100 percent ripple current. I talked with the designer of this capacitor line and he says their internal design specification is that the capacitor self heating will be less than 5 deg C from the ripple current and 100 percent voltage.

So this capacitor rating in this application will be 1790 X 0.99 or 1772 ma rms at 25 khz. This capacitor is 10 by 16 MM so will fit fine. Life should be at least 4000 X 2 (voltage) X 22.7 (case temp of 60 deg C) - or 181000 hours .

The 35 volt version EEUFM1V471 is a 5000 base hour capacitor at 105 deg C and 100 percent voltage rated at 2180 ma rms ripple at 100 khz. This would be 2180 X 0.99 or 2158 ma at 25 khz. LiIfe would be 5000 x 2 x 22.7 or 227000 hours. Again I was told that life is determined solely by voltage and case temp as long as ripple current is less than ratings.

Since I do not have a cost issue , I am going to order this 35 volt Panasonic cap and replace the Ill Caps 477KXM 35M with it the next time I have the Orion open.

I was told TenTec service supplied (Oct 2007) a capacitor from XICON as a repair replacement for C6 , C7. Why C6 , C7 instead of C6 , C9 - I do not understand - C7 runs the coolest and has the lowest ripple of the three capacitors. The supplied capacitor does not have enough marking to clearly identify its series beyond it is a 470 ufd 16 volt 105 deg C capacitor . From those markings and its size it is apparently from their ESRL low esr series. If so it is a ESRL16V470. The rated ripple current at 105C and 100 percent voltage is 635 ma. The multiplier for ripple current is 1.0 for 25 khz and approximately 1.85 for 60 deg C case temp (they only give 105 C as 1 and 85 deg C as 1.7 on the data sheet ) Thus the operational current rating would be 635 x 1.85 x 1 or 1174 ma rms.

While it is an improvement over the original 477CKR016M it is still well below the actual 1530 ma ripple current in C6. I was unable to get any more detailed Engineering info on the XICON capacitor either from XICON or Mouser. XICON just distributes the capacitor , it is made offshore and is supposed to meet JISC-5141 but they will not reveal the manufacturer to me.

Hank K7HP


QR Code
QR Code k7hpdistboardanalysis (generated for current page)