ade in the U.S. This test comes closest to measuring the modems' performance as most people would use them in the U.S.
During these tests, the modems transfer three different types of files -- compressed, graphics, and text -- ranging in size from 131 to 333 KB
, at the fastest rate possible for both one- and two-way transmissions. The large file sizes push a modem's compression engine, because not all the data can be cached in the engine's on-board dictionary.
Under a one-way transmission, modem A calls modem B, sends a file, and hangs up. Modem A then repeats the process until consistent results are achieved. Modem B only receives data; it does not send anything back to modem A. During the two-way tests, modem A still calls modem B. But when modem B answers, both devices simultaneously send files to each other.
The impaired-line tests simulate tougher-than-normal line conditions and do the best job of separating the best-designed modems from the rest. We used three lines resembling those described in the International Telecommunication Union (ITU) Telecommunications Systems Bulletin (TSB) 37A, and two more from the ITU V.56bis specification for intercontinental data transmission, ratified in August 1995. To pass, each modem must successfully make four conn
ections in six tries for each file and line condition. Of the two modems that failed to complete the test, Motorola's VoiceSURFR failed on two international lines, and Digicom Systems' ScoutPro on one.
The network-model tests measure the speed at which a modem can transfer data over the line conditions specified in Table 6 of bulletin TSB 37A. This document, created by the ITU, defines lines that model the U.S. telephone network. The 36 lines in Table 6 represent 90.55 percent of all calls made in the continental U.S.
Briefly, the procedure detailed in TSB 38 requires modem A to call modem B and send a defined, moderately compressible file that's approximately 32 KB long. To transmit approximately 300 KB, modem A sends the file 10 times before hanging up.
We also tested interoperability by connecting every test modem with every other one (see the figure
"Interoperability Test Results"
), but we didn't include the interoperability results in our product ratings. Poor interoper
ability, which is common with a new standard, is typically corrected quickly.
While every modem worked well with another of its own model, few worked as well with other models. Often, a hybrid combination resulted in data speeds somewhat slower than 28.8 Kbps, but there were two downright failures. The Archtek SmartLink 3334AV proved to be the most compatible, the Boca Research V34 33.6 the least. Interoperability is not necessarily symmetrical. For example, the Best Data SmartOne 33.6 FLX established faster connections when calling than it did when receiving.
Evaluations in this report represent the judgment of BYTE editors, based on tests conducted by NSTL, Inc., as documented in a recent issue of their monthly
PC Digest
. To purchase a copy of the full report, contact NSTL at 625 Ridge Pike, Conshohocken, PA 19428; (610) 941-9600; editors@nstl.com. For a subscription, call (800) 328-2776. BYTE magazine and NSTL are both operating
illustration_link (24 Kbytes)
We found that the first group of 33.6-Kbps modems generally work together, but not often at full speed.
Test Specs
