• Please review our updated Terms and Rules here

RLV12 Bus Termination

Agent86

Member
Joined
Aug 28, 2021
Messages
12
Location
Beech Island, SC
My PDP 11/23 has an M8061 RLV12 RL01/RL02 disk controller, but no disk drive. I've reviewed some excellent drive simulators (Thank you Reinhard) in this forum. Due to component availability, I started laying out a new interface board from from the controller's cable to a commercially available computing platform. Bus termination raised some questions.

The M8061 PCB includes twelve 82 Ohm, 1/2 Watt resistors which I'm inferring are terminators. One for each of the 12 differential signals. If driven at 5 Volts, the power would be V*V/R = 0.305 Watts.

My M8061 has three Texas Instruments SN75113N Dual Differential Line Drivers and three DEC75107B Dual Line Receivers. The 75113 data sheets show a single termination resistor between the differential lines at each end of the bus (transmission line). Termination resistance is called out as Zo, which is meant to match the impedance of the transmission line.

I don't have a DEC cable or terminator and was wondering if anyone has a schematic of the terminator? I'm thinking that the standard DEC cable includes twisted pairs for each of the differential signals. The PDP uses a ribbon cable to attach the M8061 to the backplane of the computer. The backplane also has a ribbon cable compatible connection. I plan to run a ribbon cable, no longer than 1 meter, from my computer to my emulator. I'm guessing that the DEC and M8061 terminators were based on twisted pair transmission line impedance.

My transmission line will probably have a different Zo than DEC's. I won't change the termination resistors on the M8061. My inclination is to mimic the M8061 termination on my PCB - one 82 Ohm resistor across each differential pair.

I know many of you have built, tested, and experimented with these interface boards. I would love to hear what you have to say about proper termination. I would hate to make a mistake which could possibly damage my M8061.
 

gslick

Veteran Member
Joined
Dec 30, 2010
Messages
2,112
Location
Seattle, WA
I don't have a DEC cable or terminator and was wondering if anyone has a schematic of the terminator?

RL11 Field Maintenance Print Set, MP00153
http://www.bitsavers.org/pdf/dec/uni...11EngrDrws.pdf

Terminator Assy, D-IA-7012293-0-0, Page 6 of the PDF

Looking at the drawing of the plug assembly, facing the even pins, with Pin 2 at the top and Pin 40 at the bottom. The common pins of either side of the termination array are Pin 40 and Pin 39, which are joined together around the bottom of the termination array. As noted in the drawing, the Pin 39 lead is cut at the termination array body and does not extend to the connector pins, only the common Pin 40 extends to the connector pins.

The termination array is 82-Ohms from the common Pin 40 and 39 to all of the remaining pins.
 

Agent86

Member
Joined
Aug 28, 2021
Messages
12
Location
Beech Island, SC
Thank you for the very informative response. The reference document is also very useful. I find this fascinating. Perhaps partly because I'm working on a differential bus noise problem at my day job.

The RL11's termination scheme, shown on page 30, seems rather odd for a differential bus. First, it pulls signals up to +5V and down to ground instead of applying a termination between the signals. Second, the termination is not symmetric. The RL11 pulls the low signals up to +5V while the terminator pulls all signals to ground. I suppose one justification is "It works". One advantage may be that when the drivers are off, the low signal is pulled up to 2.5V while the high signal is pulled to ground.

It also seems odd that the terminator has to cut its connection to pin UU (39). The RL11 selectively applies +5V to UU via Q2. Q2 is ON when P1_ACLO_IN_H is low. Perhaps this signal tells the drive to park its heads because the controller has lost AC power. I suspect the drive also monitors its own AC power.

Why wouldn't DEC have assigned P1_ACLO_IN_H to a different pin? Perhaps they initially wanted to symmetrically terminate the bus - pull low signals high and high signals low- then changed their mind for some reason. Perhaps they assigned the pins before designing a symmetric terminator, and it was cheaper to use 2 of the same terminators and cut a pin than to have two different terminator assemblies which could be mixed up.

I've found some 82 Ohm resistor arrays in DIP packages, but not SIP.
 

gslick

Veteran Member
Joined
Dec 30, 2010
Messages
2,112
Location
Seattle, WA
I've found some 82 Ohm resistor arrays in DIP packages, but not SIP.

Mouser has some Bourns 82-Ohm bussed array SIP parts in stock

4606X-101-820LF 6 pins / 5 resistors
4608X-101-820LF 8 pins / 7 resistors
4609X-101-820LF 9 pins / 8 resistors
4610X-101-820LF 10 pins / 9 resistors
 

Agent86

Member
Joined
Aug 28, 2021
Messages
12
Location
Beech Island, SC
DEC's RLV11/12 to RL01/02 termination unraveled

DEC's RLV11/12 uses one IC pin to drive each line of a differential pair. They should have used two IC pins.

The Dual Differential Line Drive IC on my RLV12 is a Texas Instruments SN75113. They name the signals of a differential pair Y (or AND) and Z (or NAND). For some unknown reason, TI uses separate pins for the outputs that Pull up (drive high) a given differential line and those that Sink (pull down) that same line. They name these pins P for Pull up and S for Sink. That results in 4 pins per differential pair - YP, YS, ZP, and ZS. The rest of the data sheet presumes that pins YP and YS are connected together, and the same for ZP and ZS - "The output stages are similar to TTL totem-pole outputs, but with the sink outputs, YS and ZS, and the corresponding active pullup terminals, YP and ZP, available on adjacent package pins.".

The RLV11/12 use only two of the output pins per drive, e.g. 1YP (pin 4) and 1ZS (pin 2) for driver 1. When the transistor driving 1YP high is OFF, then its line floats. When the Sink transistor driving 1ZS low is OFF, then its line floats. To transmit information reliably, you need two controlled states for each signal - high voltage and low voltage. DEC used 82 Ohm pullup and pull down resistors. They pull 1ZS high and 1YP low.

When the 1ZS transistor is ON, then the power in the pullup resistor is 0.3 Watts. DEC uses a 1/2 W resistor. DEC uses a resistor pack with 82 Ohm, 1/8 W resistors for the pull downs. 1YP attempts to drive its output ot 5 Volts. How can the 1/8 W resistor survive? The answer is that the 75113 drive includes "Short-Circuit Protection". A 9 Ohm resistor in series with the pull up transistor generates a voltage which turns on another transistor which reduces base current to the pullup transistor. It's not clear how much current the 75113 will supply to an 82 Ohm load. To keep the power below 0.125 W the current would have to be below 39 milliamps. The voltage would be 3.2 Volts. When terminated with another 1/8 W, 82 Ohm resistor to ground, the resistors could withstand a total of 78 mA. It appears TI intended current limiting to be used for protection from overloads, not as part of normal operation. It's not surprising that one post mentioned how their line drivers got hot during heavy bus use.

DEC's terminator pulls all signals down with 82 Ohm resistors. Lines that the RLV11/12 pulls up are biased to 2.5 volts. So we end up with differential signals that range from 0 to 2.5 Volts for one line and 0 to 3.2 Volts or less for the other. If DEC had used the drivers as the data sheet shows then the signals would range from 0 to 5 Volts.

I am tempted to "Fix" my RLV12 by removing the pull up and pull down resistors, tying the driver pins together, and adding appropriate termination - either on board or off. In the interest of preservation, I won't. However, if I do venture to build an RL01/02 emulator, then perhaps I will design a different termination and driver scheme for it.

"Soul of a New Machine" by Tracy Kidder describes how young engineers with little experience designed circuits quickly. Upon completion, they often realized that their designs could be improved. Management, in their haste to produce products, disallowed improvements to circuits that worked. Perhaps that is what happened here.
 

ftcnet

Experienced Member
Joined
May 4, 2019
Messages
120
Location
BC Canada
"... young engineers with little experience designed circuits quickly. Upon completion, they often realized that their designs could be improved. Management, in their haste to produce products, disallowed improvements to circuits that worked."

That's essentially true. The joke among DEC field service techs was that every new engineer that DEC hired was assigned a power supply as their first project. Once they proved incompetent at that, they could then move on to working on something more interesting.
 
Top