perry@well.UUCP (Perry S. Kivolowitz) (06/27/86)
Since there has been a lot of discussion of memory boards in this newsgroup
it is important that this be the forum for some truths about memory products
for the Amiga.
(1) Power Availablility Is Limited To ONE AMP.
I have spoken to many memory board manufacturers who have said:
``Our boards use the Amiga's power, so there is no need for a
costly power supply or messy wires.''
``Our boards can be chained together. For example, four of our
two megabyte boards will give you eight megabytes. Again
without messy wires or external power.''
``Our boards consume <some number, usually MORE THAN 500 milli
amps> of power during normal use.''
These manufacturers are stealing your money and slowly destroying your
machine. There is a HARD limit on the current draw off the side of the
Amiga. Exceeding the limit will in time lead to premature death of your
machine.
Consult the hardware spec, consult AMIGA. The power tolerances off the side
is VERY tight.
BEFORE YOU BUY A NON-SELF-POWERED BOARD FOR THE 86 PIN CONNECTOR - FIND OUT
WHAT THE POWER CONSUMPTION OF THE BOARD IS!
The GUILTY parties of this FRAUD are the LARGEST NAMES IN MEMORY BOARDS SO
WATCH OUT!!!
(2) All FAST-RAM Boards Are Not FAST.
Again, manufacturers are misleading the public. Due to their lame
designs their boards incur excessive WAIT states. WAIT states are periods
in which processing is delayed due to memory hardware overhead.
A competent designer will produce a memory board without WAIT
states.
IT COSTS YOU NO MONEY TO ASK ``HOW MANY WAIT STATES DOES YOUR BOARD
INCUR AND HOW OFTEN?''
(3) Memory Boards Are Particularly Prone To NOISE
And still again, these same manufacturer's are cutting corners with
respect to the quality of their boards. Memory boads should definately have
a ground plane to reduce susceptability to noise.
If the memory board you're looking at doesn't have one, dont buy it.
The ONLY memory products I've seen that pass ALL of the tests are:
o The mother board memory on the Byte-By-Byte expansion
chassis.
o The Convertible(TM) series of memory boards by ASDG, Inc.
(General Availablility: Late August).
Perry S. Kivolowitz
Author of Miga-Mania and Co-Chairman of JAUGrichr@pogo.UUCP (Rich Rodgers) (06/30/86)
Usenet is an excellent medium for the expression of information and the exchange of ideas. Unfortunately, many posters see the opportunity to make their views known to thousands of people as a tacit authorization to make sweeping accusations about things they may or may not have first-hand knowlege of. Most people are sensitive enough to qualify their remarks, limmiting them to the extent of their personal experience. Unfortunately, there are those who post thinly-vieled attempts to promote their opinions or interests under the guise of "insider information". The relative anonymity of the net does not, I believe, absolve one of the responsibility of journalistic ethics, particularly when addressing topics of wide public interest. Many of the remarks below could be quite damaging to a reputable manufacturer whose products did not happen to get reviewed by the original author for inclusion on his "accetable products" list. This is the worst kind of journalistic distortion, one in which there is an implied condemnation of people or things that don't serve to the interests of the writer. Below, I offer my specific critisisms. In article <1353@well.UUCP> perry@well.UUCP (Perry S. Kivolowitz) writes: >Since there has been a lot of discussion of memory boards in this newsgroup >it is important that this be the forum for some truths about memory products >for the Amiga. > >I have spoken to many memory board manufacturers who have said: (followed by purported quotes) > >These manufacturers are stealing your money and slowly destroying your >machine... Sounds like a calm, unbiased reaction to me. Are the manufacturers whose boards draw more than half an amp the same ones who say you can use more than one board at a time? Who did you talk to when you spoke to these manufacturers? What manufacturers didn't you talk to? > > >The GUILTY parties of this FRAUD are the LARGEST NAMES IN MEMORY BOARDS SO >WATCH OUT!!! > We would hope that factual data could be made available to the net as soon as it becomes available. Comments such as the above are of very little benefit to readers on the whole, and are an indictment of innocent companies. In addition, they tend to denigrate the generally excellent level of discussion found in this newsgroup. >(2) All FAST-RAM Boards Are Not FAST. > > Again, manufacturers are misleading the public. Due to their lame >designs their boards incur excessive WAIT states.... > >(3) Memory Boards Are Particularly Prone To NOISE > > And still again, these same manufacturer's are cutting corners with >respect to the quality of their boards.... Who? Who? Who? To be sure, these are valid concerns, but not neccessarily applicable to all manufacturers not listed below. > >The ONLY memory products I've seen that pass ALL of the tests are: > > o The mother board memory on the Byte-By-Byte expansion > chassis. > o The Convertible(TM) series of memory boards by ASDG, Inc. > (General Availablility: Late August). > . >Perry S. Kivolowitz /|\ >Author of Miga-Mania and Co-Chairman of JAUG | | Mmmmm... NO Disclaimer???? ____________________| -- Rich Rodgers tektronix!pogo!richr
kellym@shark.UUCP (Kelly McArthur) (06/30/86)
In article <1353@well.UUCP> perry@well.UUCP (Perry S. Kivolowitz) writes: >Since there has been a lot of discussion of memory boards in this newsgroup >it is important that this be the forum for some truths about memory products >for the Amiga. > I certainly support any efforts to disseminate information about new Amiga products, but I feel that the vindictive tone of this article was a bit much. Sounds like Mr. Kivolowitz himself may have gotten burned on some un-named memory product. Having previously posted some remarks on CardCo's aMEGA board, I feel inclined to offer the following information: >(1) Power Availablility Is Limited To ONE AMP. >..... I measured a *Production* board as drawing ~390 mA. It is low because of a CMOS controller chip, not because it uses expensive CMOS ram chips. In addition, the board layout has provisions for an external supply. I know that a few pre-production boards with non-CMOS controllers and two-layer ECB's were floating around back during Comdex. I can't imagine anyone thinking that one of these was representative of a production unit, but who knows? I suppose that could exlain Perry's concern... >(2) All FAST-RAM Boards Are Not FAST. .....[some boards induce excessive wait states] Well, I didn't happen to look at it on a logic analyzer, but I didn't notice that programs executed slower out of "fast" ram than out of "chip" ram. Remember that occasional wait states will be introduced on any board to arbitrate between refresh and access cycles. >(3) Memory Boards Are Particularly Prone To NOISE >.....[EMI, etc] The aMEGA card has both a four-layer board and a metal housing. My understanding is that it has passed FCC. >The ONLY memory products I've seen that pass ALL of the tests are: > > o The mother board memory on the Byte-By-Byte expansion > chassis. > o The Convertible(TM) series of memory boards by ASDG, Inc. > (General Availablility: Late August). > This is the really nasty part of the message. To use Mr. Kivolowitz's notation, I hope that HE has in fact looked at ALL the products that are available and is NOT making SWEEPING GENERALIZATIONS about manufacturers that have not gone out of their way to obtain the "Kivolowitz-Seal-Of-Approval" Kelly McArthur {declabs...ihnp4...hplabs}!tektronix!shark!kellym
wagner@utcs.UUCP (07/01/86)
Is it possible to talk specifics here? I would like to know which of Perry's 'rules' each of the currently available memory boards breaks. I think we need some informed, detailed comment here. Lots of us are on the verge of buying, and we need all the information we can get. Michael Wagner (wagner@utcs)
mykes@3comvax.UUCP (Mike Schwartz) (07/02/86)
>>> some writing about Amega board from CardCo and Flames about what >>> Perry Kivlowitz (sorry if I misspelled it)... I just got my 1MB board from CardCo, and it looks wonderful! It is supposed to be 100% Amiga approved, passes through the bus, fits nicely on the side of the Amiga. After booting and copying Manx library to ram: I have 1.29 MB free memory. It autoconfigs fine, and all my software appears to run fine. The only negatives I have seen so far have nothing to do with the CardCo board. Autoconfig requires 80K of ram for I don't know what (maybe someone who knows can fill us in). Also, the AddBuffers command in 1.2 uses chip memory exclusively, so it is not possible to add as many as 750 buffers (I wish I had 2MB and could add 1788 buffers to each drive). Another bit of news about CardCo... They went out of business a week or two ago. When I called to order the ram board, the phone was answered by Computers Anonymous, a retail computer store. But not to fret, apparently Computers Anonymous is indeed a "phoenix" (rises from the ashes) company from CardCo, and there will be no shortage of CardCo boards, probably ever. CardCo has a fine reputation with its C64 product line, and their first Amiga product looks wonderful. I do not have any flames for Perry, as I think he would have mentioned CardCo IF HE HAD KNOWN about them or it must have slipped his mind. He did not mention any companies with bad boards, only those that he knew were good ones. I'm sure it is safe to say, he will add CardCo to his "approved" list. I appreciate the fact that Perry was simply warning us to beware. I would hope that if someone gets burned, we all would find out so we don't get burned ourselves. Slander, on the other hand, is against the law and you can get sued for it, so proper judgement should be used when flaming products or companies. However, Ralph Nader doesn't get sued too much, does he?
perry@well.UUCP (Perry S. Kivolowitz) (07/08/86)
In article <567@3comvax.UUCP>, mykes@3comvax.UUCP (Mike Schwartz) writes: > know what (maybe someone who knows can fill us in). Also, the AddBuffers > command in 1.2 uses chip memory exclusively, so it is not possible to add > as many as 750 buffers (I wish I had 2MB and could add 1788 buffers to This is because DMA is controlled done on the internal (512K limited) bus. > I do not have any flames for Perry, as I think he would have mentioned > CardCo IF HE HAD KNOWN about them or it must have slipped his mind. He did > not mention any companies with bad boards, only those that he knew were > good ones. I'm sure it is safe to say, he will add CardCo to his > "approved" list. I appreciate the fact that Perry was simply warning us > to beware. I would hope that if someone gets burned, we all would find > out so we don't get burned ourselves. Damn straight, however, I am not maintaining an ``approved'' list. How am I to do so? You hit it right on the head - I'm simply making people aware that all memory boards are not created equal. Is the CardCo card multilayered? If you hold it up to the light can you see (easily) through it? In my original article I argue that a ground plane is important. Someone argued that a well designed grid of power and ground lines is good enough. Well, fine - those who agree with me can be memory boards with ground planes and KNOW what they get. Those who don't can hope that the board they bought (without a ground plane) was ``well designed.'' What does CardCo suggest as to how many cards can be added after the CardCo RAM card? Related question: how much current does the board draw? Are there any by-pass capacitors on the board? By-Pass cap supress noise but drive the cost of the board (more holes to drill) up (plus the cost of the caps). Many would argue, however, that having them (one or two per ram chip) is worth while. Perry S. Kivolowitz
mykes@3comvax.UUCP (Mike Schwartz) (07/09/86)
In article <1400@well.UUCP> perry@well.UUCP (Perry S. Kivolowitz) writes: >In article <567@3comvax.UUCP>, mykes@3comvax.UUCP (Mike Schwartz) writes: >> know what (maybe someone who knows can fill us in). Also, the AddBuffers >> command in 1.2 uses chip memory exclusively, so it is not possible to add >> as many as 750 buffers (I wish I had 2MB and could add 1788 buffers to > >This is because DMA is controlled done on the internal (512K limited) bus. > Obviously the dma for the disk drive works in only the lower 512K of memory, but that does not preclude moving the buffers to fast memory with the CPU. When you have ram and floppy disks (i.e. no hard disk), backing up the ram disk becomes the biggest problem facing you. Right now, I use MAKE to back up sources and objects as it makes the program, but it would be ideal if I could simply add 750 buffers to each disk and use these buffers as a cache. My development habits would not have to change with/without the extra RAM, everything that needed to be in RAM would end up there incrementally as I do things. Anything I expect to be written to disk ends up on disk automatically. 750 buffers is not unreasonable, when you consider that that is < 440K. 440K taken away from my chip memory is unreasonable. So, I guess the bottom line is that the AddBuffers command is a great first step toward making the machine faster with floppies, but it still doesn't do exactly what I would like it to do (granted, it is only my wish, I have not seen anyone else request it yet). So for now, I use 30 buffers for both drives and have to hack at my makefile to get the desired effect. This is ok, but it could be friendlier, in my humble opinion. >What does CardCo suggest as to how many cards can be added after the CardCo >RAM card? Related question: how much current does the board draw? Are there >any by-pass capacitors on the board? By-Pass cap supress noise but drive the >cost of the board (more holes to drill) up (plus the cost of the caps). Many >would argue, however, that having them (one or two per ram chip) is worth >while. > >Perry S. Kivolowitz Apparently, the Cardco boards draw 390 MilliAmps, so two boards can be plugged in at one time. They do pass through the bus, and the only problems I have had with it are the logistics of using it efficiently and reliably. I haven't taken it apart, but I do have high confidence in CardCo, since I have seen LOTS of their products and they ALL have been correct. I would recommend the board to anyone who is looking for RAM. The cost is fair, and it is a well done product. It is supposedly the first board that does follow the Amiga add-on specs. It doesn't appear to have the ability to be plugged into a card cage, but since it passes the bus, a card cage can still be used with it.
cmcmanis@sun.uucp (Chuck McManis) (07/09/86)
On the subject of RAM cards. The Allegra from Access Associates is four
layer, generates practically *no* RFI with the case *removed* and none
at all with it on. All ram has bypass caps, but it does *not* pass the
bus. Simple measurements on the Vcc line show my board draws 350mA for
512K.
--Chuck
{Bay Area}!sun!cmcmanis All these opinions are my own.daveh@cbmvax.cbm.UUCP (Dave Haynie) (07/09/86)
> What does CardCo suggest as to how many cards can be added after the CardCo > RAM card? Related question: how much current does the board draw? Are there > any by-pass capacitors on the board? By-Pass cap supress noise but drive the > cost of the board (more holes to drill) up (plus the cost of the caps). Many > would argue, however, that having them (one or two per ram chip) is worth > while. > > Perry S. Kivolowitz Last I heard, CardCo was claiming that 2 boards could be used without exceeding 1 Amp. This was based on the use of some CMOS parts, and probably refreshing several RAM banks at separate time to keep down the peak power requirements. I can't imagine anyone producting a RAM board (or any computer board, for that matter) without bypass caps. The extra costs fo PC board is negligable, at least when you've got a machine making all your boards, as most anyone marketing large quantities of a product will. At least one bypass cap per chip is good, though often a few more will be required to pass FCC requirements. Since the CardCo board has a bus feedthrough, its FCC requirements will be more stringent, since they'll have to pass with something connected to the extender. The cost of good 0.1uF or 0.22uF bypass caps these days is about 2 cents a piece or less, so for a 1 Meg board you'd pay around 80 cents or so. I sure HOPE any manufacturer considers this mandatory; if they don't, they deserve doom. -- /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh "I don't feel safe in this world no more, I don't want to die in a nuclear war, I want to sail away to a distant shore And live like an ape man." -The Kinks These opinions are my own, though for a small fee they be yours too. \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
jimm@amiga.UUCP (James D. Mackraz) (07/10/86)
Keywords:
::
Let me first say that I haven't had the opportunity to verify the
following, but you can test for yourself.
We have had reports that the drhystone benchmark runs significantly
slower in extended memory than in a 512K machine, if that extended
memory is CardCo. Those with CardCo mem, can you check? It would
be interesting to measure differences in copying large files from
RAM:a to RAM:b for a 512K machine vs. CardCo extended.
One guy here claims to have hooked up a scope and observed a "goodly
number" of wait states in the CardCo card. Can anyone verify?
Flame this spot -> X if your observations are contrary to this second-
hand information.
go wild.
jimm
"I have my two megabytes, what's your problem?"perry@well.UUCP (07/15/86)
In article <574@3comvax.UUCP>, mykes@3comvax.UUCP (Mike Schwartz) writes: > to have the ability to be plugged into a card cage, but since it passes > the bus, a card cage can still be used with it. Wrong! This is exactly what I am on about. The manufacturer would happily have you connecting a card cage to their memory board but check the AMIGA hardware spec. It basically says that results of passing the bus more than once cannot be predicted (due to line lengths, buffering delays, and the like). Serious manufacturer's will not play with your money and your head in this way (fudging specifications). Maybe CardCo could produce quality for the C64. The Amiga demands a different level of expertise. Perry S. Kivolowitz