New Production S-100 Systen Enclosures?

[Tor]

Linear vs. switched.. I noticed Grant Stockly went for switchmode in his Altair replicas, the only major difference from the original: http://www.altairkit.com/diffcompare.html#5._Power_Supply

-Tor

 

[Chuck(G)]

Yeah, it only makes sense. If SMPSUs had been widely available in 1974, the S100 bus would likely not have used regulators on individual boards. I remember Bob Boschert's operation in the mid 70's. SMPSU design was as much black art then as it was engineering. In some respects, it still is. An awful lot of energy gets wasted in the traditional S100 setup that typically uses linear regulators on each board. If you assume that the +5 power supply rail furnishes +8 per spec, then 37.5 percent of power is wasted as heat in the regulators alone. What with rectifier and transformer losses, I'd be surprised if total power supply loss wasn't somewhere around 50% in the typical S100 system.

Multibus, which is contemporary with S100, uses regulated supplies--it was (lots) more expensive in terms of front-end costs.

 

[RickNel]

An awful lot of energy gets wasted in the traditional S100 setup that typically uses linear regulators on each board. If you assume that the +5 power supply rail furnishes +8 per spec, then 37.5 percent of power is wasted as heat in the regulators alone.

.. and the heat load from those on-board Voltage Regulators is often a long-term source of stress on nearby parts, such as the associated smoothing capacitors and the solder joints around them. Cooler boards will last longer. But that's for efficiency, if you prefer that to "authenticity". These days not many S100 systems are maintained for their functional value or efficiency.

 

[Chuck(G)]

.. and the heat load from those on-board Voltage Regulators is often a long-term source of stress on nearby parts, such as the associated smoothing capacitors and the solder joints around them. Cooler boards will last longer. But that's for efficiency, if you prefer that to "authenticity". These days not many S100 systems are maintained for their functional value or efficiency.

Well, there do exist 3-terminal switching regulators, which could take a lot of the load off the power supply and lower the heat quite a bit. I remember that a 7805 regulator was insufficient to carry the load of the original MITS 4K DRAM board--they had to bridge it with a resistor. So it was a flawed idea right out of the gate.

 

[tingo]

I remember Bob Boschert's operation in the mid 70's. SMPSU design was as much black art then as it was engineering.

Oh, that bring back memories. I my first job (this was in the late 80's) I worked as an electronics repair technician for the Army in my country. Repaired all sorts of telephone, communication and computer equipment.
Once I got a challenge; fixing a large number of communication boxes (sorry, can't provide details) which all had the same problem: no power. Nobody else had dared to try on this challenge.

Said boxes had a SMPSU which used several 741 op amps for both the switching and regulator part. The darn thing was so sensitive that if you put an oscilloscope probe or a multimeter probe in the wrong place, you blew at least a couple of op amps, and most likely some more components too.

I took me about a week (and a large number of blown op amps) to get down a set of measurements that allowed us to find out which components were bad, without blowing up any more components.
For some reason, the manufacturer hadn't provided a fault-finding guide (we had service manuals and full schematics of course) for this equipment.

 

[MicrocomputerSolutions]

I had hoped to get Integrand Research to agree to drag out the spec's for one of their former products and build them for us, but since they're out of business, I'm going to have to come up with another plan. Another Enclosure and Power Supply Company has taken over production of a limited number of Integrand's last products, and they indicated a willingness to consider producing a custom enclosure for us. I asked if they had the drawing and spec's for Integrand's former S-100 product line, and I was told that I would have to talk to an engineer.

Looking at what's available in prefab knockdown enclosures, I really don't see anything that I'd be happy with. One of the key problems is the the prefab stuff is not rated to support the weight of a full size S-100 System, not even taking in the idea of the additional weight of some drive bays.

So far, while some people have made interesting suggestions, I haven't seen many people come forward to say that they'd be willing to pay real money for new production System and Drive Enclosures. From what I see, this is going to be a long term project.

I'm going to explore the enclosure manufacturing business and see what it would cost to have something made. I'm going to carry a Compupro System enclosure and Integrand Research Enclosure as examples of what I would find acceptable to see what it would cost, and how many we would have to commit to get a manufacturer to agree to producing them.

In the meantime, I'm going to look for a source of linear power supplies that could be mounted in the footprint provided in Compupro System Enclosures, so I can complete and sell the last two Compupro System Enclosures that I have left, and possible get spares so I can service Compupro power supplies in the future.

 

[MicrocomputerSolutions]

I'm not convinced that using a or some switching power supplies to replace the linear power supply is the best idea. Being in the PC sales and service business, I see the results of failed switching power supplies on a regular basis. It's not pretty. If the power supply shuts itself down like it's suppose to to protect the system, there's no problem. But too frequently, before a switching power supply fails completely, it sends out some destructive trash on the voltage outputs, often trashing the motherboard, drives, and expansion cards in the PC.

With the value of legacy S-100 board rising as the number of surviving and operable boards and systems shrinks, I'm not sure that it's a good idea to build timebombs into new enclosures. Sure, many old linear power supplies are needing service now, and will in the future, but they are easily and economically repaired to new condition.

I've seen the damage that a defective power supply (switchers and linear) can do to systems when they fail, and I'll put my faith in a old Compupro linear power supply over a modern switcher/switchers made in the Far East any day.

 

[smp]

I had hoped to get Integrand Research to agree to drag out the spec's for one of their former products and build them for us, but since they're out of business, I'm going to have to come up with another plan.

Rats! That's too bad.

So far, while some people have made interesting suggestions, I haven't seen many people come forward to say that they'd be willing to pay real money for new production System and Drive Enclosures. From what I see, this is going to be a long term project.

Well, speaking only for myself, I already have an IMSAI chassis, as well as an Integrand chassis, so I am not under much pressure to acquire another S-100 chassis at present. Perhaps that is the same case with many folks here? I certainly remain interested in the outcome of your effort, but I just wanted to offer that bit of feedback.

I'm going to explore the enclosure manufacturing business and see what it would cost to have something made. I'm going to carry a Compupro System enclosure and Integrand Research Enclosure as examples of what I would find acceptable to see what it would cost, and how many we would have to commit to get a manufacturer to agree to producing them.

I agree that is an excellent plan to move forward. Good luck, and please do keep us posted.

smp

 

[atod]

I would like an enclosure that's like the California Computer Systems (CCS) model. Simple and small. Hopefully inexpensive as well.